Method and arrangement for identifying a ground fault in a drive unit of a vehicle

Abstract

A ground fault in a vehicle drive unit comprising an inverter is detected. The input side of the inverter is connected to a floating DC supply voltage and is controlled by a control device 12. The con

Description

Technical Field The invention relates to a method with which a ground fault in a floating system, in particular in a drive unit of a vehicle, can be identified and where required located, and to a corresponding arrangement. Background A floating system is sometimes used to supply electricity in vehicles, in particular rail vehicles. A floating system (also referred to as an insulated system or IT system) is distinguished in that all live conductors are insulated from ground. The term "insulated" here also encompasses a resistive connection to ground. In an IT system, in the event of a single ground fault (i.e. a low-impedance connection of only one conductor to ground), operation is still possible since a high residual current usually does not occur. In a vehicle, in particular a rail vehicle, ground faults can occur around the engines, for example due to damage to the cable insulation. At such a point, there is a high voltage ripple with respect to the supplying DC voltage system since the engines are connected to the supply system via inverters, for example pulsewidth-modulation inverters. In the event of a ground fault on the AC voltage side of the inverter, this ripple influences the supply system (DC voltage side). This can lead to components on the DC voltage side, such as interference suppressor filters for example, being overloaded, to impermissible EMC interference or interference currents in the grounding system, to resonant overvoltage and / or to further damage throughout the system. A single ground fault must therefore be quickly detected and located and then the relevant sub-system must be disconnected if possible. However, reliably detecting a single ground fault is difficult and finding the cause of the fault is complex, in particular in the case of an extensive system with a large number of loads, or in the case of non-permanent or irregularly recurring ground faults. It is known to use insulation monitoring units which measure the impedance between the energized system and the ground potential. Furthermore, it is known to evaluate voltage peaks occurring in the event of a ground fault, or to evaluate a shift in potential of the floating system. The cause of the ground fault is often found by successively switching on the loads in a lengthy process, or a residual current detector is used. It is therefore desirable to provide a method and a corresponding arrangement with which a ground fault in a drive unit of a vehicle can be reliably detected and located. Summary The present invention is defined by the independent claims. Advantageous configurations of the invention are specified in the dependent claims. The invention is based on the knowledge that a ground fault on the AC voltage side (output) of an inverter in a drive unit of a vehicle leads to large changes on the DC voltage side (input) of the inverter, i.e. the change over time dU / dt of the input voltage of the inverter to ground (common mode) is significantly greater than during interference-free operation. During interference-free operation, the voltage on the input side of the inverter is usually not absolutely constant but rather has spectral components originating from the supply system, such as harmonic frequencies, for example, which arise during conversion in substations. These are, however, characterized by a significantly lower change over time dU / dt than the spectral components caused by a ground fault. The change over time dU / dt is checked by a checking unit connected at the input and is compared with a predefined value. If this value is exceeded, the checking unit generates a control signal and provides it to a control unit for controlling the inverter. Provision can be made for the checking unit to generate the control signal only if the value is exceeded several times within a determined time interval. The control signal preferably terminates the operation of the inverter. The control unit can be configured such that it passes operating commands which are required for normal operation to the inverter only if an enable signal is present, and the control signal blocks the enable signal. Alternatively or in addition, in the event of a fault (if dU / dt is thus greater than in the interference-free case), the checking unit can send a control signal to the control unit, which control signal ensures that the control unit or the control device sends a switch-off command to the inverter and / or sends no operating commands which are required for normal operation to the inverter. Suitable parametrization of the components ensures that only the inverter of the relevant drive unit is switched off. The checking unit measures the input voltage, which is applied to the inverter, of the drive system to ground. In one embodiment, the checking unit has a filter in the form of a high-pass filter which lets through the high frequencies, that is to say the spectral components of the signal which typically occur in the event of a ground fault, and which suppresses the lower frequencies, that is to say the spectral components during normal operation. The spectral components of the voltage that are let through are routed to an evaluation unit which ascertains the change over time dU / dt. This evaluation can be performed in an analog or digital manner. 27 08 24 5 The checking unit can additionally have a low-pass filter which, together with the high-pass filter, constitutes a bandpass filter. This is particularly advantageous if the subsequent evaluation in the evaluation unit is performed digitally; the evaluation unit thus comprises an analog-to-digital converter. 10 The bandpass filter then also acts as an anti-aliasing filter for the digital processing. A digital evaluation unit comprises, in addition to an analog-to-digital converter, a processor unit for determining dU / dt and 15 for forming the control signal. The processor unit can for example measure the digitized values of the voltage U(t) at different times tl, t2, form the difference between them and compare this difference with a predefined threshold value. If the threshold value is exceeded, the control signal can be 20 generated. However, it is also possible for each exceedance of the threshold value to be counted in a counter and for the control signal to be generated if the number of exceedances in a predefined period exceeds a predefined value. Alternatively or in addition to a processor unit, the digital evaluation unit 25 can have a logic circuit, an FPGA, a microcontroller or the like . An analog evaluation unit can comprise a rectifier and a comparator. The comparator compares the rectified voltage with 30 a predefined threshold value and directly or indirectly generates the control signal. The analog evaluation unit can additionally comprise a device for smoothing the rectified voltage so that a sufficiently smoothed voltage is applied to the input of the comparator. In this case, owing to the high-35 pass filter, no voltage or only a low voltage is applied during interference-free operation. According to another embodiment, no frequency filtering takes place at the input of the checking unit. The voltage, measured by the checking device, at the input of the inverter is digitized 5 and the checking unit determines the change over time dU / dt thereof. The checking device can have a digital evaluation unit as described above. The embodiment can be realized purely by software . 10 Brief description of the drawings The invention will be explained in more detail below with reference to exemplary embodiments through the use of the accompanying drawings in which: 15 Fig. 1 shows an exemplary embodiment of the invention; Fig. 2 shows an exemplary embodiment of the control device; Fig. 3 shows an exemplary embodiment of the processor unit; and Fig. 4 shows an exemplary control device. 20 27 08 24 Detailed description Fig. 1 shows an example of a drive system of a rail vehicle. It comprises a drive unit 1; further drive units are denoted with 25 la, lb and can be of identical construction. The drive units are supplied with power from the rail network 2, 3 via busbars or overhead lines; the housings of components are grounded via wheels 4 of the vehicle. The rail network 2, 3 is operated as an insulated system and provides a DC voltage as the supply 30 voltage (DC supply voltage) . In this case, the two lines 2, 3 are at a potential of for example +400 V and -250 V, respectively. The drive unit 1 comprises a motor 5 which is connected to an 35 inverter 7 via lines 6. The inputs 8, 9 of the inverter are connected to the DC supply voltage 2, 3, wherein an inductor 10 for decoupling from the further drive units la, lb is present in both connection lines. The drive unit 1 can be disconnected from the DC supply voltage by way of switches 11. The drive unit 1, in particular the inverter 7, is controlled by a control device 12 the construction and mode of operation of which are explained in more detail in the following figures. The control device 12 receives commands 13 from a superordinate system via a first interface and sends operational commands 14 to the inverter 7 via a second interface, by means of which operational commands the inverter is controlled. The control device 12 further has a connection 15 to the supply potential, in particular the positive supply potential 2, and a connection to ground 16. Furthermore, there is an interference suppressor filter 17 connected between the positive supply potential 2, the negative supply potential 3 and ground, that is to say housing. The interference suppressor filter 17 is used to guide clockfrequency currents occurring during interference-free operation (i.e. in the clock cycle of the inverter, for example 1450 kHz) back to the converter 7 so that they do not enter into the supply system. Such clock-frequency currents can arise due to parasitic capacitances of components on the AC current side, in particular due to engine cables. The interference suppressor filter 17 can be formed in the conventional way and in particular have capacitors and resistors. Fig. 2 shows the control device 12 in detail. The (positive) input voltage of the inverter to ground is measured via the two connections 15 and 16. The control device 12 contains a checking unit 20 and a control unit 21. In the checking unit 20, the input voltage is filtered by a bandpass filter 22 which has appropriately dimensioned capacitors 22a, 22b and resistors 22c, 22d such that spectral components of the voltage which occur during interference-free operation are damped to a sufficient extent. Conversely, the bandpass filter lets through frequencies which typically occur in the event of interference in the form of a ground fault on the AC voltage side, essentially undamped. When a bandpass filter is used instead of a single high-pass filter, in the case of subsequent digital processing an aliasing effect is simultaneously avoided since very high frequencies are not passed on. The correspondingly filtered input voltage is routed to an evaluation unit 24 via a connection 23. In this exemplary embodiment, the evaluation unit 24 comprises an analog-to-digital converter 25 which digitizes the input voltage U(t) and passes the digital values 26 on to a processor unit 27 in the evaluation unit 24. The processor unit ascertains the change over time dU / dt and compares it with a predefined threshold value. If the threshold value is exceeded or is exceeded too often, a control signal 28 is sent to the control unit. The control signal controls the inverter. For example, the control signal can trigger the sending of a switch-off command to the inverter. In another embodiment, if the threshold value is exceeded or is exceeded too often, an enable signal which is applied to the control unit 21 during interference-free operation can be disabled. Fig. 3 shows an example of the evaluation of the digitized U(t) signal 26 by way of the processor unit 27. The values U(t) are shifted into a memory, in particular into a FIFO shift register 30, wherein the clocking is for example 20 ps . At suitable times tl, t2, at least the associated values U(tl), U(t2) are taken (symbolized by 31); in the example shown, four values U(tla), U(tlb), U(t2a), U(t2b) are taken in total, wherein two values directly follow one another in each case and their mean value (32, 33) is formed. The mean value U(tl) is thus formed from U(tla) and U(tlb), wherein the time difference between U(tla) and U(tlb) is 20 ps. The same applies for U(t2a) and U(t2b) and their mean value U(t2). Forming the mean value is optional and is used for smoothing. The times tl and t2 are preferably chosen 5 such that their difference corresponds approximately to half the expected period duration of U(t) in the event of interference. In the example considered, this temporal difference is 80 ps. The difference between U(tl) and U(t2) is then ascertained (34) 10 and compared (35) with a predefined threshold value. Any exceedance of the threshold value is counted in a counter 36 which is set back to zero after a determined time. If a predefined number of threshold value exceedances is counted in this time interval, the control signal is sent to the control 15 unit. The threshold value to choose depends on the composition and dimensioning of the drive system and of the drive unit, and on the clocking of the shift register. In this exemplary embodiment, the threshold value can be 20 A / ps and the control signal can be generated if the threshold value is exceeded 20 20 times within a time interval of 50 msec. 27 08 24 Fig. 4 shows another control device 12 which can optionally also be constructed in a purely analog manner. The (positive) input voltage of the inverter 7 to ground is measured via the two 25 connections 15 and 16. The control device 12 contains a checking unit 20 and a control unit 21. In this example, the checking unit 20 comprises a high-pass filter having suitably dimensioned components (resistor 22e, 30 capacitor 22f) . The input voltage is filtered by the high-pass filter 22e, f such that spectral components of the voltage which occur during interference-free operation are damped to a sufficient extent, and frequencies which typically occur in the event of interference in the form of a ground fault pass through 35 essentially undamped. If the signal processing is performed in an analog manner, an additional low-pass filter or a bandpass filter is not necessary. The evaluation unit 24 has rectifiers 40a, 40b rectifying the switching edges. Advantageously, there is also a circuit 41 for smoothing; this is achieved by a capacitor 41a and a resistor 41b. The smoothed voltage is compared with a predefined threshold value in a comparator 42. For this purpose, use can also optionally be made of a relay, a voltage converter or an analog-to-digital converter. If the voltage applied to the comparator input is greater than the predefined threshold value, a control signal 28 is generated and sent to the control unit 21. As described above, this control signal controls the inverter; for example an enable signal applied to the control unit is disabled. The invention is based on the evaluation of the switching edges, i.e., of the change in voltage of the DC supply voltage of an inverter. If a ground fault has occurred on the output side, this change is very significant and dU / dt is very large. Through appropriate parametrization and optionally suitable filtering and through sufficient decoupling of drive units from one another and from the supply system by way of the inductors 10, good selectivity can be achieved, i.e. a ground fault on the AC voltage side can be easily distinguished from other interferences in the supply system or from a ground fault at another location. By decoupling various drive units from one another, the ground fault can be located since only the inverter of the relevant unit is accordingly controlled, preferably switched off. No complex additional components are required for carrying out the method and so the costs are low. In the case of a purely digital evaluation, only corresponding software components are required, i.e., the processor unit must be able to perform the corresponding software steps. In the case of other embodiments, only a few, inexpensive passive elements, such as the components 22 for frequency filtering, need to be provided. The inductors 10 for decoupling are usually already present, as is the interference suppressor filter 17 which is used to smooth 5 harmonics in the DC supply voltage. This interference suppressor filter can also house the filter components 22, for example. The features and aspects of the invention described in the exemplary embodiments can of course be combined with one another 10 in various ways. In particular, the features can be used not only in the described combinations but also in other combinations or by themselves.

Claims

27 08 241. A method for detecting a ground fault in a drive unit of a rail vehicle, wherein5 the rail vehicle comprises a plurality of drive units eachof which is supplied with power from a rail network that is operated as an insulated system, the rail network comprising two busbars to which a negative voltage and a positive voltage of a floating DC supply voltage are applied respectively,10 each drive unit comprises an inverter, wherein the inverteris connected on its input side to the floating DC supply voltage via inductors and is controlled by a control device,the control device is connected to the supply voltage and comprises a checking unit and a control unit,15 the checking unit comprises a high-pass filter or abandpass filter for filtering the voltage at an input of the inverter, and an evaluation unit connected to the high-pass filter or the bandpass filter,the evaluation unit comprises an analog-to-digital20 converter and a processor unit, wherein the analog-to-digital converter digitizes the filtered voltage and passes the resulting digital values to the processor unit,the processor unit evaluates a change over time dU / dt of the voltage at the input of the inverter based on the digital 25 values,the processor unit provides a control signal to the control unit if the change over time dU / dt exceeds a predefined value several times within a predefined time interval, andthe control device uses the control signal as a basis for 30 controlling the inverter.

2. The method of claim 1, wherein the control signal switches off the inverter.3527 08 243. An apparatus for detecting a ground fault in a drive unit of a rail vehicle, comprising:a plurality of drive units, wherein each drive unit is configured to be supplied with power from a rail network that 5 is operated as an insulated system and that comprises two busbars to which a negative voltage and a positive voltage of a floating DC supply voltage are applied respectively, and wherein each drive unit comprises an inverter which is configured to be connected on its input side to the floating DC supply voltage 10 via inductors, anda control device for controlling the inverter, wherein the control device is connected to the supply voltage and comprises a checking unit and a control unit, whereinthe checking unit comprises a high-pass filter or a 15 bandpass filter for filtering the voltage at an input of the inverter, and an evaluation unit connected to the high-pass filter or the bandpass filter,the evaluation unit comprises an analog-to-digital converter and a processor unit, the analog-to-digital converter 20 being configured to digitize the filtered voltage and pass the resulting digital values to the processor unit,the processor unit is configured to evaluate a change over time dU / dt of the voltage at the input of the inverter based on the digital values, and to provide a control signal to the 25 control unit if the change over time dU / dt exceeds a predefined value several times within a predefined time interval, andthe control device is configured to use the control signal as a basis for controlling the inverter.

Images