INTERSTRAND SHORT CIRCUIT TESTING OF STATOR WINDING BARS OF ELECTRIC MACHINES

Abstract

A method and an apparatus check a multiplicity of mutually insulated strands in a stator winding bar of an electric machine. The method includes injecting a test signal, determining a first component of the test signal, and comparing at least the first component with a reference signal in order to determine damage to insulation between individual strands of the multiplicity of strands. The apparatus includes a signal source and a measuring apparatus. The method and the apparatus are particularly suitable for checking stator winding bars as are used in a generator and/or an electric motor.

Claims

1. A method for checking a multiplicity of mutually insulated strands in a stator winding bar of an electrical machine, the method comprising: injecting a test signal into the multiplicity of strands of the stator winding bar, wherein the step of injection comprises an injection at a first segment of the stator winding bar or at a first end segment of the stator winding bar, detecting a first component of the test signal, wherein the step of detection comprises a detection of a first reflection component at the first segment or at the first end segment, injecting the test signal into at least a second selection of the multiplicity of strands of the stator winding bar at a second segment of the stator winding bar or at a second end segment of the stator winding bar, detecting a second reflection component at the second segment or the second end segment, comparing at least the first component of the test signal with a reference signal in order to detect damage to an insulation between individual strands of the multiplicity of strands, wherein the reference signal is a superposition reference signal that is compared with a superposition signal of the first reflection component and the test signal and is with a superposition signal of the second reflection component and the test signal.

2. The method as claimed in claim 1, further comprising: terminating a second end segment of the stator winding bar by a short circuit, a resistor or the characteristic impedance of the stator winding bar.

3. The method as claimed in claim 2, wherein the step of termination comprises a termination of at least a first selection of the multiplicity of strands, preferably the multiplicity of strands at the second end of the stator winding bar.

4. The method as claimed in claim 1, wherein the test signal comprises a temporal voltage curve with at least one edge of a defined gradient.

5. The method as claimed in claim 1, wherein the test signal comprises a temporal voltage curve with at least one voltage pulse with at least one edge of defined gradient.

6. The method as claimed in claim 1, wherein the defined gradient amounts to a few volts within a few nanoseconds.

7. The method as claimed in claim 1, wherein the test signal is at least one temporal voltage curve with at least one maximum of a few volts.

8. An apparatus for checking a multiplicity of strands in a stator winding bar of an electrical machine, the apparatus comprising: a signal source configured to provide and inject a test signal into the multiplicity of strands, a measuring apparatus configured to detect a reflection component and/or a transmission component of the test signal at a first segment or at a first end segment of the stator winding bar, and a second reflection component at a second segment or at a second end segment of the stator winding bar, and to analyze the reflection component and/or the transmission component in relation to a reference signal, wherein the reference signal is a superposition signal, and is with a superposition signal of the second reflection component and the test signal.

9. The apparatus as claimed in claim 8, wherein the reference signal is a superposition reference signal, and is with a superposition signal of the first reflection component and the test signal.

10. The apparatus as claimed in claim 8, wherein the electrical machine is a generator.

11. The method as claimed in claim 6, wherein the defined gradient amounts to 0.1 to 10 volts per nanosecond or higher.

12. The method as claimed in claim 7, wherein the test signal is at least one temporal voltage curve with at least one maximum of a few volts in the range from 0.1 to 80 volts.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Preferred embodiments of the method according to the invention and of the apparatus according to the invention are explained below with reference to the attached schematic drawing. Here:

[0024] FIG. 1a shows a first arrangement for carrying out the method according to the invention,

[0025] FIG. 1b shows a second arrangement for carrying out the method according to the invention,

[0026] FIG. 2a shows an exemplary plot of reference signals Uref and reflection components in the case of a superposition signal U1+U2,

[0027] FIG. 2b shows a schematic signal plot of the reference signal Uref for the reflection component in the case of an undamaged insulation and of a faulty insulation, and

[0028] FIG. 3 shows a flow diagram of the method according to the invention.

DETAILED DESCRIPTION OF INVENTION

[0029] FIG. 1 shows a first structure for carrying out the method according to the invention. For this purpose, a stator winding bar 1 is subjected at a first segment, in particular at a first end segment 1a, to a test signal U1, there being no termination or short-circuit provided at a second segment of the stator winding bar 1, in particular at its second end segment 1b. The test signal U1 can be provided by a signal source 12. The test signal U1 propagates along the stator winding bar 1. It is advantageous for the test signal U1 to be injected into a multiplicity of the strands within the stator winding bar 1, in order to be able to check a plurality of the strands with one measurement. It is possible, without limitation, for the test signal U1 to be injected into all of the strands of the stator winding bar 1.

[0030] Suitable means for the injection 100 of the test signal as the first step of the method are known to the expert. The injection 100 can thus signify, for example, an electrical connection, a capacitive injection and/or an inductive injection of the test signal U1. In FIG. 1a, the first end segment 1a of the stator winding bar 1 is shown as the segment for the injection 100. The expert will, nevertheless, understand that, without limitation, the injection 100 of the test signal U1 into the multiplicity of strands can also be done at any other segment of the stator winding bar 1. The location of the injection can therefore be selected in accordance with the nature of the specific fabrication. Due to the characteristic impedance of the stator winding bar 1, the test signal U1 will propagate along the stator winding bar 1, wherein a transmission component U3 reaches the second end 1b of the stator winding bar 1, while, due to the given characteristic impedance, a reflection component U2 is reflected to the injection segment 1a—shown in the figure as the first end segment of the stator winding bar 1.

[0031] If now, as a result of damage to a mutual insulation of the strands of the stator winding bar 1, low-resistance connections or short-circuits between individual instances of the strands of the stator winding bar 1 are present, then the characteristic impedance of the stator winding bar 1 changes at these fault locations. The change to the characteristic impedance of the stator winding bar 1 also entails a change in the reflection component U2 and in the transmission component U3. When the injection 100 of the reference signal U1 occurs at the first end segment 1a or the first segment of the stator winding bar 1, the reflection component U2 measured there can also be referred to as a first reflection component. The first reflection component U2 is detected or measured in a step of detection 200. The significant change of this first reflection component U2 in the event of a low-resistance connection between individual instances of the strands can be detected in a comparison step 300 of the method. For this purpose, the first reflection component U2 is compared with a reference signal Uref for the fault-free case, in order to detect damage to the insulation between individual instances of the strands. A measuring apparatus 14 can, for example, be used for this purpose, perhaps in the form of an oscilloscope.

[0032] The expert will understand that, on the basis of the significant change in the characteristic impedance in the case of a fault, a significant change to a transmission component U3 of the test signal U1 can also be used for the detection of damage to the insulation, i.e. low-resistance connections between individual instances of the strands. The measuring apparatus 14 could also be attached to the second end segment 1b of the stator winding bar 1 (not illustrated) to measure the transmission component U3.

[0033] FIG. 1b shows an alternative structure for carrying out the method according to the invention. In contrast to FIG. 1a, the second end segment 1b of the stator winding bar 1 is terminated by a resistor Z, Z0. The use of the resistor has the advantage that in the case of an undamaged insulation between the strands, a defined reflection behavior of the test signal U1 can be set up along the strands. Through the resistor Z, Z0 for example, a short-circuit to ground, or a termination by a characteristic impedance Z0 of the correctly mutually insulated strands can be achieved.

[0034] For the structure for carrying out the method according to the invention illustrated in FIGS. 1a and 1b, the test signal U1 is only injected at a first end 1a of the stator winding bar 1. It is of course possible, without limitation, in addition or alternatively to inject the test signal U1 at the second end segment 1b of the stator winding bar 1. At the second end segment 1b, the test signal U1 is injected into at least one second selection of the strands of the stator winding bar 1.

[0035] It is advantageous for the injection 100a of the test signal U1 to be made into the same strands that were already used for the injection 100 at the first segment or end segment 1a. If the measurements for the respective reflection components U2, U2′ for these two measurements differ, so that the first reflection component U2 at the first segment or end segment 1a is different from a second reflection component U2 at the second segment or end segment 1b of the stator winding bar 1, this is an indication of damage to the strands. This means that in a step of comparison 300a of the second reflection component U2′ with a reference signal Uref and/or the first reflection component U2, a significant difference emerges that points to damage of the mutual insulation of the strands.

[0036] For the measuring set up illustrated in FIG. 1a and FIG. 1b, a test signal U1 of a few volts, with a defined edge gradient, is sufficient. The edge gradient should be in the range of a few volts within a few microseconds. Equally, it is possible, without further limitation, for the gradient to be in the range of a few volts per nanosecond. As is known, the sensitivity for the detection of fault locations rises with increasing gradient; this is because a broader frequency spectrum is contained in the edge as the gradient increases.

[0037] The expert will understand that in the case of a detection of the first reflection component U2 at the first end segment 1a, and of the second reflection component U2′ (not illustrated) at the second end segment 1b of the stator winding bar 1, it is possible that an occurrence of damage to the insulation that is symmetrical with respect to the length of the stator winding bar 1 can be overlooked in the step of comparison 300a.

[0038] FIG. 2a illustrates by way of example a plot of a superposition signal U1+U2 against time t. A characteristic signal Uref arises, as illustrated, in the case of an undamaged mutual insulation of the strands. As a result of the significant change of the reflection component U2 in the case of a damaged insulation between individual instances of the strands, a significant change occurs to the reflection component U2, both at the first end segment 1a or at the second end segment 1b of the stator winding bar 1.

[0039] It is advantageous to plot a superposition signal U1+U2 against time, since in the event of a fault, the superposition signal U1+U2 decreases, thus resulting in a significant deviation between the reference signal Uref and the superposition signal U1+U2. Further options for superposition signals are known to the expert. Without limitation it is, for example, possible for the reflection component U2, U2′ detected in step 200 to be enlarged or amplified in order to make the deviation yet clearer. It is therefore possible to select suitable representations, such that a significant difference between the fault-free stator winding bar and a faulty stator winding bar results in the step of comparison 300a.

[0040] FIG. 2b shows, by way of example, a plot of the first or second reflection signal U2, U2′ against time. As a result of the significant change to the characteristic impedance in the event of a damaged insulation, a significant difference (as shown) arises between the reference signal Uref and the first or second reflection component U2, U2′. This significant difference can be used to detect damage to the insulation between individual instances of the strands in the step of comparison 300, 300a of a desired reference signal with the respective reflection components U2, U2′.

[0041] It is also without limitation possible, in the step of detection 200, to use the measured transmission component U3 at the respectively opposite segment or end segment of the stator winding bar 1, which is not used for injection, for detection of damage to the mutual insulation of the strands of the stator winding bar, in the place of the first reflection component U2 or of the second reflection component U2′. The significant change to the characteristic impedance in the case of a low-resistance connection between individual strands also has an effect on the transmission component U3.

[0042] Further possibilities are known to the expert for generating a superposition signal from test signal U1, its reflection component U2, U2′ and/or the transmission component U3 in order to achieve a significant difference at the step of detection 300, 300a between an undamaged insulation and a damaged insulation. Illustrations of selected signals from the test signal U1, the reflection component U2, U2′ and/or the transmission component U3 against one another in the form of Lissajous figures are conceivable, for example. The signal curves shown in FIGS. 2a and 2b are to be understood as merely exemplary, and on no account as restricting the method of the invention or the apparatus 10.

[0043] Methods are moreover known to the expert for preparing templates from signal curves to be expected for the case of a fault and the fault-free case, so that it may be possible to distinguish between the two cases automatically.

[0044] FIG. 3 shows a flow diagram of the method according to the invention for checking a multiplicity of mutually insulated strands of the stator winding bar 1. In a step 100 of the injection, a test signal U1 is injected into the plurality of strands of the stator winding bar 1. The multiplicity of strands can here comprise all of the strands present in the stator winding bar 1; it is equally possible, without limitation, that an injection in step 100 is only made into a selection of the mutually insulated strands of the stator winding bar 1. The step 100 can, alternatively or in addition, comprise a step of injection 100a of the test signal into at least a second selection of the multiplicity of strands of the stator winding bar at a second segment or at a second end segment 1b of the stator winding bar 1.

[0045] The method can, further, advantageously comprise a step 110 of the termination of a second end segment of the stator winding bar 1. The termination can be made through a short circuit, a resistor Z or the characteristic impedance Z0 of the stator winding bar 1. Typically the termination 110 is made at the end segment of the stator winding bar 1 that is not used for injection 100 of the test signal. As described, the step of termination changes the reflection behavior for the correctly mutually insulated strands. It can be advantageous for all the strands into which the test signal is injected in the step 100 also to be terminated in step 110.

[0046] According to FIG. 3 the method further comprises a step of detection 200 of a component of the test signal U1. The detection 200 can involve the detection 200 of the first reflection component U2 and/or of the transmission component U3 of the test signal U1. Equally it is optionally possible that the second reflection component U2′ of the test signal U1 is detected in a step 200a.

[0047] In a step 300 at least the first reflection component U2 and/or the transmission component U3 is compared with a reference signal Uref in order to detect damage to an insulation between individual strands of the multiplicity of strands. Optionally, the step of comparison 300 can also comprise a comparison 300a between at least the second reflection component U2′ and the reference signal Uref and/or the first reflection component U2. The steps 300, 300a of the method can comprise the possibilities of comparison described in connection with FIGS. 2a and 2b.

[0048] The apparatus 10 according to the invention for checking a multiplicity of strands in a stator winding bar 1 of an electrical machine is illustrated in FIGS. 1a and 1b.

[0049] The apparatus 10 comprises a voltage source 12 for the provision and injection of a test signal U1 into the plurality of strands of a stator winding bar 1. The apparatus 10 further comprises a measuring apparatus 14. The measuring apparatus 14 can, for example, be an oscilloscope. The measuring apparatus 14 serves to detect a reflection component U2, a second reflection component U2′ and/or a transmission component U3 of the test signal U1. The measuring apparatus 14 can further permit an analysis of the reflection components U2 and/or U2′ in relation to a reference signal Uref. The reference signal Uref here typically refers to a corresponding signal for a stator winding bar 1 whose insulation between the individual strands is undamaged. The apparatus 10 can further comprise means for injecting the test signal U1 into the multiplicity of strands. The appropriate means are known to the expert, and are not illustrated in FIGS. 1a and 1b. The voltage source 12 and the measuring apparatus 14 can also, without limitation, be implemented as one apparatus 10. Optionally the apparatus 10 can further comprise a resistor Z, Z0 for terminating the stator winding bar. The reference signal Uref can, without limitation, be the forms of the reference signal described above in connection with FIG. 2a and FIG. 2b.