Partial discharge detection

Abstract

The invention relates to a method and a device (1) for detecting partial discharges at the windings of an electrical machine. According to the invention, a first electromagnetic measurement signal is detected in the partial discharge frequency range and a second electromagnetic measurement signal is detected in the interference frequency range. Furthermore, a time window around the first measurement signal is predetermined and a recognition criterion for the partial discharge signal and an interference signal criterion for the second measurement signal are defined. If a first measurement signal is detected that meets the recognition criterion and no second measurement signal that meets the interference signal criterion is detected within the time window around the first measurement signal, a partial discharge is inferred.

Description

Technical Field

[0001] The invention relates to detecting partial discharges at the windings of an electric machine. Background Art

[0002] Partial discharges are local electrical discharges caused by strong inhomogeneities of the electric field, for example in local inhomogeneities in the insulating material, in particular in small air enclosures in the insulation of an electrical conductor. Partial discharges can locally damage the insulation and lead to long-term insulation failure. Therefore, electric machines are checked for partial discharges on a regular basis. Partial discharges cause short electromagnetic pulses with a typical duration of less than 1 μs and frequency components up to the UHF frequency range. Usually, partial discharge signals are searched for in the frequency range between approximately 100 kHz and 10 MHz, in which the partial discharge pulses are usually the strongest.

[0003] There are a large number of partially standardized measuring methods for detecting partial discharges, for example in accordance with DIN EN 60270. Almost exclusively offline measurements are performed on electric machines, wherein the normal operation of the electric machine is interrupted and high costs are incurred, for example due to operational interruptions and costs for measuring equipment. Such measurements are therefore only performed occasionally.

[0004] From US 2007 057677 A1 a device and a method for finding partial discharges in a conductor bar of a dynamoelectric machine are known, wherein the conductor bar is provided with an external electrical insulation. A first and a second sensor are configured to detect a signal generated by the partial discharge, which propagates along the conductor bar. The first and the second sensor are arranged at intervals at the conductor bar in order to emit a partial discharge output signal representative of a specific detection point. The partial discharge output signals of the first and the second sensor are supplied to an evaluation unit, which is configured such that the evaluation unit localizes the partial discharge on the conductor bar.

[0005] A monitoring device for monitoring the degree of activation of partial discharges in an AC power system is known from US2018120380A1. The device has: a pair of high-frequency and low-frequency sensors; a time control module arranged to receive a cyclic signal from the low-frequency sensor and provide a time control signal derived from the timing of a continuous cycle of the cyclic signal; and a monitoring unit arranged to digitize the signal from the high-frequency sensor, detect the digitized signal as continuous data blocks of different durations, and trigger the detection of each data block in response to a clock signal. An electrical system with state monitoring includes a speed-adjustable drive and a motor driven by the drive, as well as a pair of sensors.

[0006] In the case of an offline measurement, the measurement setup can be performed in a limited environment and external interference factors are usually (but not always) largely excluded. In contrast, in the case of an online measurement during normal operation of the motor, all operating components are active and can affect the measurement by coupling in interference signals. If, for example, a rotating motor is operated at a converter, the partial discharge pulses are superimposed with a large number of interference signals caused by the switching process of the converter. For example, radio services, such as radio broadcasting, professional radio stations or mobile radio, non-specific electromagnetic compatibility (EMC) interference sources, unapproved radio equipment, equipment working in the industrial, scientific and medical (ISM) frequency band (e.g. high-frequency welding equipment, laser equipment), radio antennas, alarm devices and / or radio switches are considered as additional interference signals. Which interference signals can appear and what frequencies these interference signals have depend on the specific characteristics of the components of the object in which the partial discharge should be detected (e.g., components of the system consisting of a converter and a motor), and on the interference sources present in the surrounding environment of the object.

[0007] In industrial facilities, interference often occurs at the converters due to adjacent machines. This can be a permanent problem. Welding work in the environment can also be a source of interference. However, in maintenance work, this is only a temporary interference, but in the case of welding robots, it is also a permanent interference. Interference signals are in particular pulsed interference with a broad frequency spectrum in the range of up to 1 MHz or 100 MHz. The frequency range here depends on the source of the interference.

[0008] In order to reliably detect partial discharges, such interference signals must be separated or distinguished from the partial discharge signal so as not to confuse the partial discharge signal with the interference signal. Summary of the invention

[0009] The invention is based on the object of improving the detection of partial discharges in windings of an electrical machine, in particular with regard to the differentiation between partial discharge signals and interference signals.

[0010] According to the invention, this object is achieved by a method having the features of the invention and by a device having the features of the invention.

[0011] In the method according to the invention for detecting partial discharges at a winding of an electrical machine, a first electromagnetic measurement signal is detected in the partial discharge frequency range and a second electromagnetic measurement signal is detected in the interference frequency range. Furthermore, a time window around the first measurement signal is predetermined and an identification criterion for the partial discharge signal and an interference signal criterion for the second measurement signal are defined.

[0012] Partial discharges can occur at different locations, such as slots or winding ends. Different types have different patterns in the phase-resolved partial discharge image. The time window can be adapted to the expected time window or can be adapted to the expected time window.

[0013] TE pulses typically have a duration of about 100 ns to less than 5 μs. In one embodiment, a time window can be placed less than 5 μs before and after the pulse.

[0014] If a first measurement signal is detected which meets the identification criterion and no second measurement signal is detected in a time window around the first measurement signal which meets the interference signal criterion, a partial discharge is inferred.

[0015] Therefore, according to the invention, in addition to the electromagnetic measurement signal for detecting partial discharges, referred to here as the first measurement signal, a plurality of further electromagnetic measurement signals, referred to here as the second measurement signals, are detected for detecting interference signals which may influence or even generate the first measurement signal.

[0016] In one embodiment, the second measurement signal can also be a signal that is not an electromagnetic signal. This can be implemented, for example, in an adjacent motor at the inverter by measuring the structure-borne sound there and converting the structure-borne sound into a signal. The inverter clock cycle is detected. Of course, the propagation time in the machine housing must be taken into account here. Structure-borne sound pulses reach the sensor much later than electrical interference at the partial discharge sensor. In another embodiment, for example, the control signal of the power semiconductor of the inverter is used directly for interference suppression, i.e. as the second measurement signal, wherein the control signal can again be an electrical signal or an optical signal.

[0017] In order to identify the first measurement signal as a candidate for a partial discharge signal, detection of the first measurement signal is restricted to a partial discharge frequency range that is typical for partial discharges and defines an identification criterion that the first measurement signal must meet in order to be classified as a candidate for a partial discharge signal. The second measurement signal is detected in an interference frequency range that is associated with an interference signal that can influence or generate the first measurement signal.

[0018] Furthermore, in particular, a disturbance signal criterion is defined, which characterizes the second measurement signal as a disturbance signal potentially relevant for detecting partial discharges. For evaluating the second measurement signal, a time window around the detected first measurement signal is predetermined, in which a search for disturbance signals relevant for detecting partial discharges is performed if the first measurement signal is classified as a candidate for a partial discharge signal by means of the identification criterion. It is thus taken into account that the first measurement signal can only be influenced or generated by disturbances which occurred shortly before the detection of the first measurement signal (in particular a few μs), so that disturbance signals which occur at least approximately simultaneously with the first measurement signal are relevant for detecting partial discharges.

[0019] In one embodiment, the size of the time window can be determined or varied. Here, for example, the distance between the antennas, the influence of the signal propagation time in the propagation medium and the time tolerance of the pulse detection (jitter) need to be taken into account. The time window needs to be selected to be larger than this factor.

[0020] In one configuration, the beginning and / or the end of the detected first measurement signal is determined via an amplitude threshold value, wherein the detection can be performed by means of a comparator.

[0021] The time window takes into account, for example, the duration which elapses between the origin and the detection of the first and / or second measurement signal and / or during the processing of the first and / or second measurement signal. Thus, according to the invention, if a first measurement signal is detected which is a candidate for classification as a partial discharge signal and no relevant interference signals are detected in the time window around the first measurement signal, a partial discharge is therefore inferred. The simultaneous evaluation of the first measurement signal and the interference signals which can potentially influence and generate the first measurement signal advantageously increases the reliability of the identification of partial discharges relative to the evaluation of only the first measurement signal. The influence relates in particular to the superposition of signals.

[0022] A design of the present invention provides that the identification criterion includes at least one feature of an amplitude signal of the amplitude of the spectral component of the first measurement signal, which is related to the frequency of the spectral component. For example, the amplitude signal is an envelope signal proportional to the course curve of the logarithmic amplitude of the spectral component of the output signal of the frequency filter. This design of the present invention advantageously simplifies the evaluation of the first measurement signal, because the partial discharge signal is a high-frequency signal, and the direct evaluation of the high-frequency signal is complex and requires a lot of computing power. Features suitable for defining the identification criterion are, for example, the pulse duration and / or the limits of the level of the amplitude signal and / or features that characterize the signal form of the amplitude signal (for example, the limits of the ratio of the level to the pulse duration). Interference signals (for example, interference signals caused by adjacent inverters) often appear periodically and with a high repetition frequency.

[0023] Therefore, another embodiment of the present invention provides that the interference signal criterion includes at least one feature of the amplitude signal of the amplitude of the spectral component of the second measurement signal which is related to the frequency of the spectral component.

[0024] For example, the repetition frequency can be used as an identification criterion. The bandwidth of the partial discharge rate in good insulation of <100 pulses / s up to damaged insulation of >100K pulses / s can be very large. Strict periodicity is a feature that indicates an interference source. This can be identified by statistical characteristics. The distribution of the time intervals of adjacent pulses can also be used as a criterion. In addition, not only direct pulses can be considered, but also multiple adjacent pulses and spectrum evaluation and / or automatic correlation, in particular automatic correlation with suspicious interference frequencies in the identification criteria.

[0025] In another embodiment of the invention, it is checked whether the second measurement signal forms a regular signal sequence of signals of the same type, for example, which are repeated at equal time intervals, and the interference signal criterion includes the dependence of the second measurement signal on such a signal sequence. This embodiment of the invention takes into account that interference signals are usually generated periodically, while partial discharge signals usually occur in a randomly distributed manner.

[0026] In a further embodiment of the invention, the partial discharge frequency range and the interference frequency range have a non-empty intersection. This advantageously takes into account in particular second measurement signals which have a frequency in the partial discharge frequency range and are therefore potentially relevant for partial discharge detection.

[0027] If the interference signal occurs periodically and / or has a specific pattern, artificial intelligence can be used in one design to identify the signal. For example, the following characteristics can be derived from the statistical evaluation of the time series, which allow the pulses to be classified as TE (partial discharge) and interference. In particular, a permanent logical link (PLL) logged on to the interference signal can be used. If a specific interference source has a characteristic pulse pattern, it can be detected by pattern recognition.

[0028] In another embodiment of the invention, at least a subset of the first measurement signals is detected in or at the insulation of the winding of the electric machine. This embodiment of the invention advantageously makes it possible to detect the first measurement signals in the vicinity of the generation of the partial discharge to be detected and thereby improve the measurement sensitivity of the partial discharge detection. In addition, the influence of interference signals on the first measurement signals is reduced, since the interference signals are usually generated outside the electric machine.

[0029] In another embodiment of the present invention, at least a subset of the second measurement signal is detected outside the motor, in particular in the vicinity of a potential source of interference with the second measurement signal. As already explained, interference signals are usually generated outside the motor. Therefore, detecting the second measurement signal outside the motor, in particular in the vicinity of a potential source of interference, is conducive to achieving a high measurement sensitivity for interference signal detection. In particular, it can be proposed that the potential source of interference is determined, located and eliminated as much as possible in advance and the second measurement signal that cannot be eliminated is detected in a targeted manner in the vicinity of the located source of interference.

[0030] The location where the interference signal or partial discharge signal is detected can play a role. The current transformer is a potential source. Therefore, it is particularly expedient to arrange the sensor for measuring the signal near the current transformer and / or near the machine supplied by the current transformer. In this case, "near" means in particular in the range of the first 15% of the cable length of the cable, in which range the measurement is performed at the cable.

[0031] In another embodiment of the present invention, a distinction criterion for distinguishing a first measurement signal from a second measurement signal is defined. Furthermore, in the case where a second measurement signal that meets the interference signal criterion is detected in a time window around the first measurement signal that meets the identification criterion, the first measurement signal is compared with the second measurement signal, and a partial discharge is inferred only if the first measurement signal meets the distinction criterion. Therefore, this embodiment of the present invention is directed to the case where the second measurement signal is detected in a time window around the first measurement signal, the first measurement signal is classified as a candidate for a partial discharge signal, and the second measurement signal is a potentially relevant interference signal for partial discharge detection. This embodiment of the present invention proposes that in this case, the first and second measurement signals are compared with each other in order to decide whether the first measurement signal is attributed to a partial discharge or to an interference source. In order to make this decision, a distinction criterion for distinguishing a first measurement signal from a second measurement signal is defined and used.

[0032] An extended design of the above-mentioned design of the present invention provides that the distinction criterion includes at least one feature, which distinguishes the amplitude signals of the amplitudes of the spectral components of the first measurement signal and the second measurement signal from each other according to the frequency of the spectral components. In addition, the extended design provides that the distinction criterion includes the lack of periodic repetition of the first measurement signal, which is specific to the second measurement signal. The first extended design provides that the first measurement signal and the second measurement signal are compared according to the comparison of the amplitude signals formed by the first measurement signal and the second measurement signal. This has the advantages of the simplified signal amplitude evaluability compared to the direct evaluation of the measurement signal mentioned above. The second extended design also makes use of the fact that interference signals are often generated periodically, while partial discharge signals usually appear in a randomly distributed manner. Therefore, if the second measurement signal repeats periodically, while the first measurement signal does not repeat periodically, it is unlikely that the first measurement signal can be attributed to the interference signal.

[0033] In a further embodiment of the invention, the interference signal criterion comprises a threshold value for the signal strength of the second measurement signal, and the evaluation of the first measurement signal is interrupted if the signal strength of the second measurement signal exceeds the threshold value. According to this embodiment of the invention, the evaluation of the first measurement signal is therefore interrupted in the event of very strong interference signals, since very strong interference signals could influence the partial discharge signal and thus call into question the validity of the evaluation of the first measurement signal. In this case, the threshold value can preferably be varied and thus in particular adapted to changing circumstances and requirements for partial discharge detection.

[0034] In another embodiment of the present invention, it can be provided that the detected second measurement signal is evaluated in order to specifically find interference sources that affect the electromagnetic quantity of the partial discharge in order to eliminate the interference sources as much as possible. It can also be provided that the partial discharge detection is evaluated in terms of its quality or reliability based on the frequency of the second measurement signal and / or at least one other characteristic value thereof (e.g., the occurrence time point of the second measurement signal, the signal strength and / or the signal duration). Thus, for example, the greater the frequency and / or strength of the second measurement signal in time correlation with the first measurement signal, the worse the quality of the partial discharge detection is classified.

[0035] The device according to the invention for detecting partial discharges at a winding of an electric machine according to the method according to the invention comprises two measuring devices and at least one evaluation unit. The first measuring device is configured to detect a first electromagnetic measurement signal. The second measuring device is configured to detect a second electromagnetic measurement signal. The at least one evaluation unit is configured to check whether the first measurement signal meets a criterion for identifying a partial discharge signal and whether a second measurement signal that meets a criterion for an interference signal is detected within a predetermined time window around the first measurement signal.

[0036] In one embodiment of the device according to the invention, the first measuring device has at least one first antenna and, for each first antenna, a coupling-out unit for coupling out a signal detected by the first antenna, wherein the first antenna is arranged in or on an insulation of a winding of the electric machine.

[0037] In another embodiment of the device according to the invention, the second measuring device has at least one second antenna and, for each second antenna, a second coupling-out unit for coupling out a signal detected by the second antenna, wherein the second antenna is arranged outside the motor, in particular in the vicinity of a potential interference source that generates the second measurement signal.

[0038] In particular, the first and / or the second measurement signal can be detected at different locations, for example by means of a plurality of antennas.

[0039] In a further embodiment of the device according to the invention, at least one evaluation unit is configured to compare the first measurement signal with the second measurement signal based on a differentiation criterion if a second measurement signal satisfying the interference criterion is detected in a time window around the first measurement signal satisfying the identification criterion.

[0040] In a further embodiment of the device according to the invention, the at least one evaluation unit is designed to interrupt the evaluation of the first measurement signal if the signal strength of the second measurement signal exceeds a threshold value.

[0041] The device according to the invention is able to carry out the method according to the invention with the advantages already mentioned above. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above-described characteristics, features and advantages of the present invention and the manner and method of achieving the same will become clearer and more distinct in conjunction with the following schematic description of the embodiments, which are explained in detail in conjunction with the accompanying drawings. Herein, it is shown:

[0043] Figure 1 A block diagram showing a first embodiment of a device for detecting partial discharges,

[0044] Figure 2 A block diagram showing a second embodiment of a device for detecting partial discharges,

[0045] Figure 3 A flow chart showing one embodiment of a method for detecting partial discharges.

[0046] Parts that correspond to one another are provided with the same reference symbols in the figures. DETAILED DESCRIPTION

[0047] Figure 1A block diagram of a first embodiment of a device 1 for detecting partial discharges at a winding of an electrical machine is shown. The device 1 comprises two measuring devices 3 , 5 and an evaluation unit 7 .

[0048] The first measuring device 3 is configured to detect an electromagnetic first measurement signal in the partial discharge frequency range. To this end, the first measuring device 3 has a first sensor unit 9 and a first frequency filter 11. The first sensor unit 9 has a first antenna 13 arranged in or at the insulation of the winding of the motor and a first coupling-out unit 15 for coupling out the signal detected by the first antenna 13. The first antenna 13 is formed, for example, by the lead of a first temperature sensor, which is configured to detect the temperature in the motor and has a temperature-dependent measuring resistor made of platinum, for example, a platinum measuring resistor Pt100. In any case, such a temperature sensor is usually used in motors and can also be advantageously used to detect partial discharges. The terminal conductors of the first antenna 13 are preferably implemented in a shielded manner to reduce the coupling-in of interference signals. The first coupling-out unit 15 is, for example, a frequency divider for coupling out the high-frequency signal detected by the first antenna 13, so as to separate the signal from the low-frequency temperature signal. The first frequency filter 11 is configured to filter the electrical signal coupled out by the first coupling-out unit 15 and has a passband that defines the partial discharge frequency range.

[0049] The second measuring device 5 is configured to detect an electromagnetic second measurement signal in the interference energy range. To this end, the second measuring device 5 has a second sensor unit 17 and a second frequency filter 19. The second sensor unit 17 has a second antenna 21 arranged outside the motor and a second coupling output unit 23 for coupling out the signal detected by the second antenna 21. The second antenna 21 is formed, for example, by a feeder of a second temperature sensor, which is configured to detect the ambient temperature in the vicinity of the motor and has a temperature-dependent measuring resistor made of, for example, platinum, for this purpose, such as a platinum measuring resistor Pt100. The terminal conductor of the second antenna 21 is preferably implemented in an unshielded manner in order to receive as many interference signals as possible. The second coupling output unit 23 is, for example, a frequency divider for coupling out the high-frequency signal detected by the second antenna 21, so as to separate the signal from the low-frequency temperature signal detected by the second temperature sensor. The second frequency filter 19 is configured to filter the electrical signal coupled out by the second coupling output unit 23, and has a passband that defines the interference frequency range. The interference frequency range and the partial discharge frequency range, for example, have a non-empty intersection. In particular, the interference frequency range can completely include the partial discharge frequency range. Alternatively, however, the interference frequency range can also be located outside the partial discharge frequency range (but preferably in the vicinity thereof), since even if the second measurement signal is detected only outside the partial discharge frequency range, a disturbance causing the second measurement signal will also cause an interference signal having a frequency within the partial discharge frequency range. Preferably, in any case, the interference frequency range is broadband than the partial discharge frequency range. The partial discharge frequency range and / or the interference frequency range can also have discontinuous local frequency ranges in order to detect partial discharges and / or interference signals in different frequency ranges.

[0050] In an alternative embodiment of the device 1, the first sensor unit 9 and / or the second sensor unit 17 have another sensor for detecting electromagnetic signals instead of the antenna 13, 21, for example, at least one capacitive coupling device with at least one coupling capacitor, an inductive coupling device with at least one coupling coil, or a directional coupler.

[0051] The evaluation unit 7 has two high-frequency detector circuits 25, 27 and a calculation unit 29. The first detector circuit 25 is configured to determine a first amplitude signal of the amplitude of the spectral component of the first measurement signal from the first measurement signal output by the first frequency filter 11 as a function of the frequency of the spectral component. The second detector circuit 27 is configured to determine a second amplitude signal of the amplitude of the spectral component of the second measurement signal from the second measurement signal output by the second frequency filter 19 as a function of the frequency of the spectral component. For example, the amplitude signals determined by the detector circuits 25, 27 are respectively envelope signals, which are proportional to the logarithmic amplitudes of the spectral components of the output signals of the corresponding frequency filters 11, 19.

[0052] The calculation unit 29 comprises two analog-to-digital converters 31 , 33 , two timers 35 , 37 , two event storage units 39 , 41 and a microcontroller 43 .

[0053] The first analog-to-digital converter 31 digitizes the first amplitude signal output by the first detector circuit 25. The second analog-to-digital converter 33 digitizes the second amplitude signal output by the second detector circuit 27.

[0054] The first timer 35 determines the time at which the first measurement signal is detected and the corresponding signal duration of the first measurement signal. The second timer 37 determines the time at which the second measurement signal is detected and the corresponding signal duration of the second measurement signal. Alternatively, instead of two timers 35, 37, a multi-channel timer can also be used, wherein the first and second measurement signals are associated with different channels of the timer.

[0055] The first event storage unit 39 stores the first amplitude signal digitized by the first analog-to-digital converter 31 and the associated time point and signal duration detected by the first timer 35. The second event storage unit 41 stores the second amplitude signal digitized by the second analog-to-digital converter 33 and the associated time point and signal duration detected by the second timer 37.

[0056] The amplitude signals digitized by the detector circuits 25, 27 are evaluated by the microcontroller 43. In particular, the microcontroller 43 checks whether the digitized first amplitude signal of the first measurement signal meets a predetermined recognition criterion for a partial discharge signal and whether a second measurement signal is detected in a predetermined time window around the first measurement signal, the digitized second amplitude signal of which meets a predetermined interference signal criterion. In addition, the microcontroller 43 is configured to compare the first measurement signal with the digitized amplitude signal of the second measurement signal if the second measurement signal is detected in a predetermined time window around the first measurement signal, wherein the digitized first amplitude signal of the first measurement signal meets the recognition criterion and the digitized second amplitude signal of the second measurement signal meets the interference signal criterion. Further aspects of the evaluation of the first and second measurement signals are described below with reference to an embodiment of the method according to the invention. Instead of the microcontroller 43, the computing unit 29 can have, for example, an FPGA (field programmable gate array) or other processing units suitable for evaluating digitized amplitude signals.

[0057] Figure 2 A block diagram of a second embodiment of a device 1 for detecting partial discharges in a winding of an electric machine is shown. Figure 1 The exemplary embodiment shown differs in that the evaluation device additionally has a delay element 49 , the delay element having a threshold value detector 47 and a switching unit 49 , and the calculation unit 29 has a digital-to-analog converter 51 instead of the second analog-to-digital converter 33 and the second event memory 41 .

[0058] The delay element 45 is connected between the first detector circuit 25 and the analog-to-digital converter 31 , and delays the forwarding of the first amplitude signal output by the first detector circuit 25 to the analog-to-digital converter 31 by a delay time τ.

[0059] The digital-to-analog converter 51 converts the threshold value of the signal strength of the second measurement signal, which is predetermined as a digital signal, into an analog threshold signal, and outputs the threshold signal to the threshold detector 47. The threshold detector 47 compares the second amplitude signal of the second measurement signal output by the second detector circuit 27 with the threshold signal. When the second amplitude signal exceeds the threshold signal, the threshold detector 47 activates the switching unit 49, which then interrupts the output of the first amplitude signal to the calculation unit 29 through the delay element 45 in the following manner, that is, the calculation unit connects the output end of the delay element to the ground potential.

[0060] In this case, the delay time τ takes into account the different signal propagation and processing times of the first and second measurement signals, in particular the time required by the threshold value detector 47 for comparing the second amplitude signal with the threshold value signal and the switching time required for activating the switching unit 49 .

[0061] For example, the threshold value can be changed via the microcontroller 43 and thus adapted to changing situations and requirements. Thus, the sensitivity of the device 1 to the strength of the interfering signal can be flexibly varied. In particular, the suppression of interfering signals can also be switched off via the threshold value in the following manner, for example by setting a very high threshold value.

[0062] In order to interrupt the evaluation of the first measurement signal by activating the switching unit 49, for example for a predetermined interruption time, the switching unit 49 is deactivated again after expiration of the interruption time, or the evaluation of the first measurement signal is continued by deactivating the switching unit 49 as soon as the second amplitude signal falls below the threshold signal again or after a predetermined duration has elapsed after the second amplitude signal falls below the threshold signal.

[0063] exist Figure 2 In the exemplary embodiment shown, only the first amplitude signal of the first measurement signal is evaluated by the microcontroller 43 by respectively checking whether the first amplitude signal of the first measurement signal meets a predetermined recognition criterion.

[0064] exist Figure 1 and Figure 2 The exemplary embodiments shown can also be combined, for example by means of delay elements 45, threshold value detectors 47, switching units 49 and digital-to-analog converters 51 being similar to Figure 2 The embodiment shown in the expansion Figure 1 The first embodiment shown. As in the second embodiment, if the second amplitude signal exceeds the threshold signal, the evaluation of the first amplitude signal is interrupted by activating the switching unit 49, but if the switching unit 49 is deactivated, that is, if the second amplitude signal is below the threshold signal, the first and second amplitude signals are processed by the analog-to-digital converters 31, 33 as in the first embodiment and evaluated by the microcontroller 43.

[0065] Figure 3 A flow chart 100 shows an exemplary embodiment of a method according to the invention for detecting partial discharges, the method having method steps 101 to 108 .

[0066] In a first method step 101 , a recognition criterion for a partial discharge signal, a disturbance signal criterion for the second measurement signal and a differentiation criterion for distinguishing the first measurement signal from the second measurement signal are defined.

[0067] The identification criterion includes, for example, at least one feature of the first amplitude signal of the amplitude of the spectral component of the first measurement signal, which is related to the frequency of the spectral component. Such a feature is, for example, a pulse duration and / or a boundary of the level of the first amplitude signal and / or a feature characterizing the signal shape (e.g., a boundary of the ratio of the level to the pulse duration).

[0068] The interference signal criterion includes a threshold value for the signal strength of the second measurement signal. In addition, the interference signal criterion can include at least one feature of the second amplitude signal of the amplitude of the spectral component of the second measurement signal, which is related to the frequency of the spectral component. Such a feature is, for example, a feature characterizing the shape of the signal and / or the level of the second amplitude signal and / or the boundary of the continuous application of the pulse, similar to the identification criterion. In addition, it can be checked whether the second measurement signal is formed with signals of the same type that repeat regularly, especially follow each other at the same time interval, and the interference signal criterion can include the properties of the second measurement signal to such a signal sequence.

[0069] The differentiation criterion comprises, for example, at least one feature which distinguishes the amplitude signals of the amplitudes of the spectral components of the first measurement signal and the second measurement signal from one another depending on the frequency of the spectral components. Alternatively or additionally, the differentiation criterion can comprise the absence of a periodic repetition of the first measurement signal as a feature of the second measurement signal. After the first method step 101, the second method step 102 is carried out.

[0070] In a second method step 102 , a first electromagnetic measurement signal is detected in the partial discharge frequency range, and a second electromagnetic measurement signal is detected in the interference frequency range in a predetermined time window around the first measurement signal. After the second method step 102 , a third method step 103 is performed.

[0071] In a third method step 103, it is checked whether the signal strength of the second measurement signal detected in the second method step 102 exceeds a threshold value. If this is the case, a fourth method step 104 is carried out after the third method step 103. Otherwise, after the third method step 103, a fifth method step 105 is carried out.

[0072] In a fourth method step 104 , the evaluation of the first measurement signal is interrupted, for example for a predetermined interruption time or until the signal strength of the second measurement signal falls below a threshold value. After the fourth method step 104 , the second method step 102 is carried out again.

[0073] In a fifth method step 105, it is checked whether the first measurement signal detected in the second method step 102 meets the identification criterion for a partial discharge signal. If this is the case, a sixth method step 106 is carried out after the fifth method step 105. Otherwise, after the fifth method step 105, the second method step 102 is carried out again.

[0074] In a sixth method step 106, it is checked whether the second measurement signal detected in the second method step 102 meets the interference signal criterion, provided that the interference signal criterion has at least one feature of the second measurement signal that is alternative to the threshold value. If the interference signal criterion does not have a feature that is alternative to the threshold value or the second measurement signal meets a feature that is alternative to the threshold value, a seventh method step 107 is carried out after the sixth method step 106. Otherwise, an eighth method step 108 is carried out after the sixth method step 106.

[0075] In a seventh method step 107, it is checked whether the criterion for distinguishing between the first measurement signal detected in the second method step 102 and the second measurement signal detected in the second method step 102 is met. If this is the case, an eighth method step 108 is carried out after the seventh method step 107. Otherwise, after the seventh method step 107, the second method step 102 is carried out again.

[0076] A partial discharge is inferred in an eighth method step 108. Furthermore, the value of a counter variable for the detected partial discharge, for example entered in the calculation unit 29, is preferably incremented by 1 and the time of the partial discharge is stored. After the eighth method step 108, the second method step 102 is carried out again.

[0077] Although the present invention has been illustrated and described in detail by means of preferred embodiments, the present invention is not limited to the disclosed examples and those skilled in the art can derive other variations therefrom without departing from the scope of protection of the present invention.

Claims

1. A method for detecting partial discharges in a winding of an electric machine, wherein - detecting a first electromagnetic measurement signal in the partial discharge frequency range, - detecting a second electromagnetic measurement signal in the interference frequency range, - predetermining a time window around said first measurement signal, - define the identification criteria for partial discharge signals, - defining an interfering signal criterion for the second measurement signal, and - a partial discharge is inferred if a first measurement signal is detected which satisfies the identification criterion and no second measurement signal is detected within the time window around the first measurement signal which satisfies the interference signal criterion, - in, The partial discharge detection is evaluated with regard to its quality, wherein the quality of the partial discharge detection is classified as worse the greater the frequency and / or intensity of the second measurement signal in temporal correlation with the first measurement signal.

2. The method according to claim 1, in, The identification criterion comprises at least one feature of an amplitude signal of an amplitude of a spectral component of the first measurement signal that is related to the frequency of the spectral component, wherein the interference signal criterion comprises at least one feature of an amplitude signal of an amplitude of a spectral component of the second measurement signal that is related to the frequency of the spectral component.

3. The method according to claim 1 or 2, in, It is checked whether the second measurement signal forms a signal sequence of regularly repeating signals of the same type, and the interfering signal criterion includes a correlation of the second measurement signal with the signal sequence.

4. The method according to claim 1 or 2, in, The partial discharge frequency range and the interference frequency range have a non-empty intersection.

5. The method according to claim 1 or 2, in, At least a subset of the first measurement signals is detected in or on an insulation of the winding of the electrical machine.

6. The method according to claim 1 or 2, in, At least a subset of the second measurement signal is detected outside the electric machine.

7. The method according to claim 6, in, At least a subset of the second measurement signals is detected in the vicinity of an interference source potentially generating the second measurement signals.

8. The method according to claim 1 or 2, in, A distinction criterion is defined for distinguishing a first measurement signal from a second measurement signal, and in the event that a second measurement signal that satisfies the interference signal criterion is detected within a time window around the first measurement signal that satisfies the identification criterion, the first measurement signal is compared with the second measurement signal, and a partial discharge is inferred only if the first measurement signal satisfies the distinction criterion.

9. The method according to claim 8, in, The differentiation criterion comprises at least one feature which distinguishes an amplitude signal of an amplitude of a spectral component of the first measurement signal from an amplitude signal of an amplitude of a spectral component of the second measurement signal from one another depending on the frequency of the spectral component.

10. The method according to claim 8, in, The differentiation criterion comprises the absence of a periodic repetition of the first measurement signal, which periodic repetition is specific to the second measurement signal.

11. The method according to claim 1 or 2, in, The interfering signal criterion comprises a threshold value for the signal strength of the second measurement signal, and the evaluation of the first measurement signal is interrupted when the signal strength of the second measurement signal exceeds the threshold value.

12. A device (1) for detecting partial discharges in a winding of an electric machine according to the method of any one of the preceding claims, the device (1) include: - a first measuring device (3) which is configured to detect a first electromagnetic measurement signal, - a second measuring device (5) which is configured to detect a second electromagnetic measurement signal, and - at least one evaluation unit (7) which is configured to check whether a first measurement signal meets a criterion for identifying a partial discharge signal and to check whether a second measurement signal which meets the interference signal criterion is detected within a predetermined time window around the first measurement signal.

13. The device (1) according to claim 12, in, The first measuring device (3) has at least one first antenna (13) arranged in or on the insulation of the winding of the electric machine and has a first coupling-out unit (15) for each first antenna (13), the first antenna being arranged in or on the insulation of the winding of the electric machine, the first coupling-out unit being used to couple out a signal detected by the first antenna (13).

14. The device (1) according to claim 12 or 13, in, The second measuring device (5) has at least one second antenna (21) which is arranged outside the electric machine and a second coupling-out unit (23) for each second antenna (21) for coupling out a signal detected by the second antenna (21).

15. The device (1) according to claim 14, in, The second antenna is arranged near an interference source that potentially generates a second measurement signal.

16. The device (1) according to claim 12 or 13, in, At least one evaluation unit (7) is configured to compare the first measurement signal with the second measurement signal based on a differentiation criterion if a second measurement signal that meets the interference signal criterion is detected within a time window around the first measurement signal that meets the identification criterion.

Images