Winding defect inspection device, manufacturing method of electric machine using the winding defect inspection device, and winding defect inspection method

Abstract

In the manufacture of a winding component performed by winding a winding (1) around a wound member (11), a sensor (2) is disposed midway along a pathway leading from one end of the winding (1) to the wound member (11), wherein: an electric discharge detecting unit (5) detects a variation in a discharge current, an amount of charge, or a voltage generated by means of a voltage applied between the sensor (2) and a defective portion of the winding (1); an accumulating unit (6) accumulates an inspection signal, which is an output from the electric discharge detecting unit (5); the inspection signal accumulated in the accumulating unit (6) is allocated to computing units (7) which calculate a time for identifying a defect position in the winding (1) by analyzing the frequency characteristics of the inspection signal.

Description

Winding defect inspection device, method for manufacturing an electric machine using the winding defect inspection device, and winding defect inspection method

[0001] The present application relates to a winding defect inspection device, a method for manufacturing an electric machine using the winding defect inspection device, and a winding defect inspection method.

[0002] If there are defects such as scratches or pinholes in the insulating layer of the windings of a rotating electric machine, the insulation between the winding and the winding core or between the winding and ground will be reduced, and insulation breakdown will occur between adjacent windings of the coils that make up the winding, which may result in reduced performance of the rotating electric machine or problems such as burnout or leakage current. Therefore, rotating electric machines require inspection of winding defects.

[0003] Conventional techniques for inspecting winding defects involve applying a voltage to electrodes arranged around the winding, detecting the discharge current or amount of charge between the winding and the electrode, and comparing the magnitude of the detected current with a threshold value to determine whether or not a defect exists (see, for example, Patent Document 1).

[0004] Furthermore, as prior art, there is a partial discharge detection device that includes a sensor unit that receives partial discharge signals and noise signals generated by application of voltage to a rotating machine, a synchronization signal generation unit that generates a synchronization signal synchronized with the cycle of the drive voltage of the rotating machine, an integration processing unit that integrates the sensor signal received by the sensor unit with the cycle of the synchronization signal to generate an integrated waveform, a partial discharge detection unit that separates partial discharge and noise from the integrated waveform of the sensor signal to calculate the amount of partial discharge, and a determination unit that determines an abnormality in the rotating machine based on the amount of partial discharge (see, for example, Patent Document 2).

[0005] JP 2009-236887 A Japanese Patent No. 6345803

[0006] According to the technology disclosed in Patent Document 1, a voltage is applied to electrodes arranged around the winding, the discharge current or amount of charge of the winding is detected while the train is running, and the magnitude of the detected signal is compared with a threshold value, thereby making it possible to determine whether or not there is a defect in the winding. However, because the discharge current or amount of charge is minute and is affected by changes in the distance between the electrode and the winding, the discharge current or amount of charge can fluctuate greatly due to the vibration of the winding while the train is running, and this has been a problem in that it is not possible to reliably determine whether or not there is a defect.

[0007] According to the technology disclosed in Patent Document 2, a sensor is placed on a stationary winding (a pre-made winding), inspection signals for multiple cycles are accumulated, and the accumulated detection signals are integrated and added together.The magnitude of the processed signal is then compared with a threshold value, making it possible to determine whether or not the winding is defective.However, the process of integrating and adding together inspection signals for multiple cycles takes a considerable amount of time, and there is a problem in that it is not possible to inspect for defects in a moving winding (for example, at 500 mm / s) in real time.

[0008] Furthermore, the techniques disclosed in the above Patent Documents 1 and 2 only allow determination of the presence or absence of a defect in the winding, and have the problem of being unable to determine the location of the defect when there is one in the winding.

[0009] The present application discloses a technique for solving the above-mentioned problems, and aims to determine the location of a defect in a winding while the train is running with high accuracy and high reliability.

[0010] The winding defect inspection device disclosed in the present application inspects defects in windings when the winding is wound around a wound member to manufacture a winding component, and includes a sensor located midway along a path from one end of the winding to the wound member, a discharge detection unit that detects a discharge current, charge amount, or fluctuations in the voltage generated by a voltage applied between the sensor and a defective portion of the winding, a storage unit that stores an inspection signal that is output from the discharge detection unit, and a calculator that allocates the inspection signal stored in the storage unit and calculates the time required to identify the defect location in the winding by analyzing the frequency characteristics of the inspection signal.

[0011] Furthermore, the method for manufacturing an electric machine using the winding defect inspection device disclosed in the present application involves manufacturing an electric machine equipped with a stator core by winding, around the stator core, windings that have been confirmed to be free of insulation defects using the winding defect inspection device.

[0012] A winding defect inspection method disclosed in the present application further comprises: a sensor disposed midway along a path from one end of the winding to a wound member; a discharge detection unit that detects a discharge current, an amount of charge, or a fluctuation in the voltage generated by a voltage applied between the sensor and a defective portion of the winding; a storage unit that stores an inspection signal that is an output from the discharge detection unit; and a plurality of computing units that perform calculations based on the inspection signals stored in the storage unit, thereby inspecting defects in the winding when manufacturing a wound component by winding the winding around the wound member, The storage unit stores the inspection signal over a certain period of time and assigns the stored inspection signal to a first computing unit, causing the first computing unit to start an operation to analyze the frequency characteristics of the inspection signal. At the same time, the storage unit starts storing the inspection signal over a second certain period of time. When the storage unit has completed storing the second inspection signal, it assigns the stored inspection signal to a second computing unit, causing the second computing unit to start an operation to analyze the frequency characteristics of the inspection signal. By repeating the above operation as many times as the number of computing units, after the inspection signal has been assigned to the last computing unit, the assignment of the inspection signal is switched back to the first computing unit, and the multiple computing units then analyze the frequency characteristics of the inspection signal to calculate the time required to identify the defect location in the winding.

[0013] The winding defect inspection device and winding defect inspection method disclosed in the present application make it possible to determine the defect position of a winding while the train is running with high accuracy and high reliability.

[0014] 8A, 8B, and 8C are simplified diagrams for explaining the principle of determining the position of a defect in a winding. A diagram showing the state of a current that flows when a discharge occurs. A diagram showing a simulation of an inspection signal. A diagram showing the analysis results of frequency characteristics. A perspective view showing a state in which a winding is wound around a wound member. A cross-sectional view showing an electric machine including a stator core that is a wound member. A perspective view showing a state in which a winding is being wound around the teeth of a stator core using a nozzle control unit. A block diagram showing the configuration of a winding defect inspection device according to Embodiment 2. A block diagram showing the configuration of a winding defect inspection device according to Embodiment 3. A flowchart showing the operation of the winding defect inspection device according to Embodiment 3. A block diagram showing the configuration of a winding defect inspection device according to Embodiment 4. Fig. 21A is a block diagram showing the configuration of a winding defect inspection device according to embodiment 5. Fig. 21B is a block diagram showing the configuration of a winding defect inspection device according to embodiment 6. Figs. 21A and 21B are diagrams showing changes in power supply voltage.

[0015] Embodiment 1. FIG. 1 is a block diagram showing the configuration of a winding defect inspection device according to embodiment 1. This embodiment will be described using an example of a defect inspection device for windings in a rotating electrical machine. The winding defect inspection device according to embodiment 1 includes a winding machine main body 14 that winds a winding 1 around a wound member 11 to manufacture a winding component for a rotating electrical machine, and a defect detection unit 9 for the winding 1. The defect detection unit 9 is composed of a sensor 2, a power supply 3, a discharge detection unit 5, a storage unit 6, and a plurality of N computing units 7, numbered 1 through Nth (N is a natural number equal to or greater than 2). Note that hereinafter, the individual computing units from the 1st through Nth computing units will be collectively referred to as computing unit 7. The sensor 2 is located midway along the path from one end 10 of the winding to the wound member 11. The power supply 3 applies a voltage between the sensor 2 and the winding 1. The discharge detection unit 5 detects a discharge current or charge amount 4 that occurs between the sensor 2 and a defective portion of the winding 1 due to a defect in the insulating layer of the winding 1 passing near the sensor 2. The storage unit 6 receives the output from the discharge detection unit 5 and stores the detection signal. The stored detection signals are assigned to a plurality of calculators 7.

[0016] The winding machine main body 14 is composed of one end portion 10 of the winding, tensioners 12a and 12b, a nozzle control unit 13, and a wound member 11. The tensioners 12a and 12b apply tension to the winding 1 from one end portion 10 to the wound member 11 to suppress vibration of the winding 1. The nozzle control unit 13 is equipped with a motor, which controls the nozzle 32 (see FIG. 12) in the x, y, and z directions shown in FIG. 1 to change the winding position of the winding 1. FIG. 2 is a block diagram showing the structure of the storage unit 6, which includes an input / output terminal 6A, a CPU (Central Processing Unit) 6B, a card 6C, an A / D converter 6D, a memory 6E, and a communication unit 6F. In FIG. 2, the input / output terminal 6A receives an analog input signal, which is converted from analog to digital by the A / D converter 6D and then stored in the memory 6E or the card 6C. The stored data is then passed to the computing unit 7 via the communication unit 6F. These operations are managed by the CPU 6B.

[0017] 3 is a block diagram showing the structure of the arithmetic unit 7, which includes an input / output terminal 7A, a CPU 7B, an A / D converter 7C, a memory 7D, and peripheral circuits 7E as its constituent elements. In FIG. 3, the input / output terminal 7A receives an input signal, which is stored in the memory 7D. The CPU 7B then performs calculations. The peripheral circuits 7E include a signal communication circuit. In the operations described below, the A / D converter 7C is not used, but it can be used as needed.

[0018] Furthermore, the defect detection unit 9 repeatedly accumulates and allocates the detected signals and analyzes the frequency characteristics to accurately calculate the defect location in the running winding 1 in real time and transmits the calculation results to the host computer 8. The winding 1 is wound while changing its position in the x, y, and z directions, and as mentioned above, motors are used to change the position in the x, y, and z directions. One role of the host computer 8 is to issue commands to the motors in the x, y, and z directions. The internal structure of the host computer 8 is made up of various calculation units, including a CPU.

[0019] Furthermore, the host machine 8 outputs a warning based on the inspection results. The host machine 8 also outputs an instruction signal to the nozzle control unit 13 to change the running of the winding 1. Here, changing the running corresponds to changing the winding position of the winding 1. As shown in FIG. 12 (described later), the host machine 8 issues a position deflection command to prevent defects from being adjacent to each other, and the motors responsible for the x, y, and z directions receive this command. Based on this, the x, y, and z positions of the nozzle 32 (see FIG. 12) change, and the winding position of the winding 1 changes. The wound member 11 is fixed, while the nozzle 32 is movable.

[0020] Next, we will explain the discharge between the winding 1 and the sensor 2. Figure 4 is an enlarged view showing the state of discharge at a defective portion of the winding. In Figure 4, one end of the winding 1, where the insulating layer 16 has been peeled off, is connected to the GND terminal of the power supply 3, and the voltage output terminal of the power supply 3 is connected to the sensor 2. When an appropriate voltage is applied to the sensor 2, if there is a defect 17 due to a scratch or pinhole in the insulating layer 16 on the surface of the winding 1, a discharge 18 will occur between the core (copper wire) 15 of the winding 1 and the sensor 2, causing a current to flow in the circuit consisting of the power supply 3, sensor 2, and winding 1. Figure 5 shows the state of the current that flows when a discharge occurs.

[0021] Next, we will explain the difficulties encountered when inspecting defects in the winding 1. As shown in Figure 1, the vibration of the winding 1 can be suppressed by applying tension to the winding 1 using tensioners 12a and 12b. However, because applying excessive tension can cause the winding 1 to break, only an appropriate amount of tension can be applied to the winding 1, and vibration of the winding 1 cannot be completely suppressed. Moreover, because the magnitude of the discharge current 19 shown in Figure 5 is quite small, even a slight vibration of the winding 1 can cause a large change in the magnitude of the discharge current 19.

[0022] For the reasons described above, an inspection based on analyzing the frequency (reciprocal of the time interval Δt) characteristics of the discharge current 19 in FIG. 5 can be highly reliable in inspecting the defective portion 17 of the winding 1. However, a common method for analyzing the frequency characteristics of the inspection signal in FIG. 5 involves collecting the inspection signal shown in FIG. 5 and calculating the frequency of the discharge current 19 using a fast Fourier transform (FFT) for analyzing frequency characteristics. Therefore, collecting data and performing the FFT takes a long time, making it impossible to inspect defects in a winding 1 traveling at high speeds in real time. For example, if the traveling speed of the winding 1 is 500 mm / s and the horizontal width of the sensor 2 (the width in the direction parallel to the traveling direction of the winding 1) is 10 mm, then inspecting the defective portion 17 of the winding 1 requires completing the FFT within 0.02 seconds, which is difficult to achieve.

[0023] Next, the principle of the defect inspection method for winding 1 according to embodiment 1 will be described. FIG. 6 is a flowchart illustrating the operation of the winding defect inspection device according to embodiment 1. Defect inspection of winding 1 begins when a voltage is applied from power supply 3 to sensor 2. First, in the discharge inspection and signal accumulation step (step S601), discharge detection unit 5 inspects for discharge between sensor 2 and winding 1. If there is no defect in winding 1, no current flows from power supply 3, and discharge detection unit 5 outputs a linear signal 20 as shown in FIG. 5 (assuming no noise is present). If there is a defect 17 in winding 1, a discharge occurs between sensor 2 and winding 1, forming a circulation circuit through power supply 3, sensor 2, and winding 1 core 15. Discharge current 19, as shown in FIG. 5, flows and is detected by discharge detection unit 5. Accumulation unit 6 continuously reads the output of discharge detection unit 5 and accumulates the inspection signal. Accumulation unit 6 may be a data logger, oscilloscope, memory high-coder, or the like.

[0024] 7 is a timing chart showing the accumulation by the accumulation unit 6 and the calculation by the calculator 7. When the accumulation unit 6 collects the test signal for a certain period of time (a fixed time) Δt1 (e.g., 5 ms), it determines that accumulation is complete (YES) (step S602) and assigns the collected test signal to the first calculator, which is the first calculator (step S603). If accumulation is not complete (NO), accumulation continues. The first calculator receives the test signal from the accumulation unit 6 over a reception time 42, begins analyzing the frequency characteristics of the test signal, and performs calculations over a calculation time 43 (step S604). At the same time, the accumulation unit 6 begins a second accumulation of the test signal over the same period of time Δt1 (step S605).

[0025] After the second accumulation of the test signal is completed, the test signal is assigned to the second second computing unit in the same manner as described above, and analysis of the test signal begins. The above operations are repeated (steps S605 to S607), and after the test signal is assigned to the final Nth computing unit (step S608), i.e., after the accumulated signal is assigned to the Nth computing unit, the accumulation unit 6 switches to assigning the test signal to the first computing unit. The number N of the first through Nth computing units is determined by the accumulation time of the accumulation unit 6 and the time required for calculation by the computing unit 7. For example, if the accumulation time of the accumulation unit 6 is 5 ms and the sum of the reception time 42 and the calculation time 43 of the first through Nth computing units are the same, 30 ms each, then Δt1×N > (reception time 42) + (calculation time 43), and therefore 30 ÷ 5 = 6, the number N of computing units 7 should be 6 or more.

[0026] Furthermore, a plurality of microcontroller units (MCUs), which are digital processors, or one or a plurality of field programmable gate arrays (FPGAs) may be used as the computing unit 7. In the calculation process of the computing unit 7, the frequency characteristics of the inspection signal are analyzed, and the defect position is calculated by taking into account the running speed of the winding 1, and the calculation results are transmitted to the host computer 8 (steps S609 and S610).

[0027] Next, the principle of determining the defect location in the winding 1 will be explained with reference to FIG. 8 . FIG. 8 is a simplified diagram for explaining the principle of determining the defect location in a winding, and is a perspective view showing the state in which the winding 1 is being wound around the wound member 11. As shown in FIG. 8A , consider the case in which the storage unit 6 stores an inspection signal for a time period Δt1 from position O to position P of the winding 1. When the portion of the winding 1 from position O to position P passes under the sensor 2, the signal from position O to position P shown in FIG. 9 is stored. If there is a defect at position Q of the winding 1, a discharge occurs between the sensor 2 and the winding 1 when position Q passes under the sensor 2, and an inspection signal 19, which is a discharge current shown in FIG. 9, is detected. Once Δt2 is determined in FIG. 9 , the defect location Q can be identified based on the running speed of the winding 1.

[0028] After the portion of the winding 1 from position O to position P passes under the sensor 2, the positional relationship between positions O and P is as shown in Figure 8B. After the inspection signal from position O to position P is accumulated as described above, the accumulation unit 6 passes the inspection signal to one of the first through Nth calculators, and upon receiving the inspection signal, calculator 7 begins calculation. As shown in Figure 8C, the sum of the reception time 42 and calculation time 43 of calculator 7 is designated as Δt3. If the traveling speed of the winding 1 is constant and the distance from sensor 2 to position d on the wound member 11 is also constant, the traveling time required for the winding 1 to travel from sensor 2 to position d is determined and designated as Δt0.

[0029] If the position of defect position Q between positions O and P can be identified, that is, if the time Δt2 shown in FIG. 9 (the time required to identify the defect position in the winding 1) can be calculated, defect position Q can be identified while the winding 1 is traveling, because the traveling speed of the winding 1 is constant. Each of the first to Nth computing units sets its own reference time 0 as the time at which an inspection signal begins to be delivered. That is, as shown in FIG. 8B , the reference time 0 of the kth computing unit is the time at which the storage unit 6 completes the kth storage and delivers the inspection signal to the kth computing unit 7 (1≦k≦N) among the first to Nth computing units. If the detection time when a defect is detected by the kth computing unit 7 of the first to Nth computing units is Δt2, the time td at which defect position Q arrives at position d on the wound member 11 can be calculated from the reference time 0 of the kth computing unit 7 (the time at which the storage unit 6 begins to deliver the inspection signal) using the following equation (1):

[0030] td = Δt0-Δt3-(Δt1-Δt2) (1)

[0031] When the first through Nth computing units detect a defect, they each communicate with the host computer 8 at the time calculated using their own calculation formula (1). For example, if there is no defect, the communication signal is set low. If a defect is detected, the communication signal is set high, and after communication, the communication signal is returned to low. In addition to the above, the time when position P of the winding 1 passes under the sensor 2 can be used as a reference (the state in FIG. 8B), and the time when defect position Q arrives at position d on the wound member 11 can be calculated as Δt0 - (Δt1 - Δt2). In short, once Δt2 is determined, the time when defect position Q arrives at position d on the wound member 11 can be determined from Δt0, since the traveling speed of the winding 1 is constant.

[0032] Next, a method for calculating the time Δt2 for identifying the defect position Q will be described. The time Δt2 for identifying the defect position Q can be calculated by analyzing the frequency characteristics of the inspection signal 19 shown in Fig. 9. As an example of a method for analyzing the frequency characteristics of the inspection signal 19, a wavelet transform according to the following equation (2) is used.

[0033]

[0034] In equation (2), f(x) is the inspection signal 19, Φ(x, a, b) is the wavelet, and the symbols a and b are the wavelet parameters. WΦ(b, a) is the conversion result, which includes the frequency components of the inspection signal 19 and the appearance time of each frequency. Next, an analysis example of the frequency characteristics of the inspection signal of equation (2) will be described with reference to FIGS. 10 and 11. FIG. 10 shows a simulation of the inspection signal. Here, the simulation signal is not an actually detected signal, but a signal created by a program based on an image of the actual signal. FIG. 11 shows the results of analyzing the frequency characteristics shown in FIG. 10 using equation (2). In FIG. 10, assume that a discharge current of 15 kHz is detected around time t = 2.5 ms. The frequency analysis results in FIG. 11 indicate that a frequency of around 15 kHz was detected at time t = 2.5 ms. It can be seen that signals with a frequency of around 15 kHz are concentrated in the meshed area in FIG. 11.

[0035] Next, we will explain the process when an actual inspection signal is detected instead of a simulated signal. As with the simulated signal, the inspection signal shown in FIG. 9 is analyzed using equation (2), and if a certain frequency (e.g., 20 MHz) is found among the resulting frequency components, it is determined that there is a defect in the winding 1. That is, as with the simulated signal shown in FIG. 11, in actual inspection, if a frequency around 20 MHz is detected, it is determined that there is a defect in the winding 1. The frequency of the actual inspected signal may be 20 MHz, 100 MHz, or 15 kHz; the timing of occurrence of the defect signal can be calculated regardless of the frequency. Δt2 is determined from the time of occurrence of the frequency (e.g., 20 MHz) (for example, Δt2 is 2.5 ms in the cases of FIGS. 10 and 11). Using this Δt2, the time td at which the defect reaches position d on the wound member 11 can be calculated, as shown in equation (1) above.

[0036] Next, we will explain an example of operation when a defect is discovered in one of the first through Nth computing units after it has completed its computation. Figure 12 is a perspective view showing the state in which a winding is being wound around a wound member. As shown in Figure 12, assume that defects 33a and 33b occur adjacent to each other in winding 1, and that the winding is wound as is and assembled into a rotating electrical machine as a component. In this case, when current is passed through winding 1, a discharge occurs between defects 33a and 33b, causing widespread breakdown in the insulation layer between defects 33a and 33b, which could potentially destroy the rotating electrical machine.

[0037] Therefore, an example of the operation when a defect is found in the winding 1 will be described using the above-described defect location identification method. Let tk1 be the time when the first communication signal indicating the defect is transmitted from the first through Nth computing units, and let tk2 be the time when the second communication signal is transmitted. The difference between these two times (Δt = tk2 - tk1) is calculated, and using the traveling speed V of the winding 1 and the vertical and horizontal widths H and W of the wound member 11, it is determined whether defect 33a and defect 33b will be adjacent when the winding 1 is wound around the wound member 11. In other words, the closer the difference Δt between the above two times is to the time it takes to wind the winding 1 around the wound member 11 once, the higher the likelihood that defect 33a and defect 33b will be adjacent.

[0038] If defects 33a and 33b are adjacent, the host computer 8 issues a warning and can change the running state of the winding 1. If a defect in the winding 1 is detected for the third time (for example, if the first computing unit detects a defect and then the third computing unit detects a defect), the difference between the time of the first and third communications from the first through Nth computing units can be calculated in the same manner as above to determine whether the first and third defects are adjacent. If a defect is detected for the Nth time, the distance 41 between the Nth defect and the first defect can be calculated by multiplying the diameter of the winding 1 by N. If this distance 41 is appropriate (a distance that does not cause discharge), the host computer 8 does not issue a warning. Furthermore, if a defect is detected for the N+1th time, it can be determined whether the defects are adjacent by comparing the detection time of the second defect.

[0039] Furthermore, to accurately calculate the time required to wind the winding 1 around the wound member 11 once, the time may be calculated by adding the vertical width H and horizontal width W shown in FIG. 12 together, taking into account the thickness of the winding layers of the winding 1 wound around the wound member 11. For example, when the diameter of the winding 1 is 0.5 mm and 10 layers of the winding 1 are wound around the wound member 11, the thickness of the winding layers is 5 mm, and this can be calculated by adding 5 mm to the vertical width H and horizontal width W. Furthermore, a correction term can be inserted into equation (1) to account for the elongation of the winding 1 due to the tension applied to the winding 1. For example, assuming that the winding 1 elongates by approximately 0.1% compared to when there is no tension, the calculation result of the following equation (3) can be added for correction.

[0040] M×0.1% / V (3) where M is the distance from the sensor 2 to the wound member 11, and V is the running speed of the winding 1.

[0041] 1 is merely one example. That is, by arranging a plurality of defect detection units 9, each including a sensor 2, it is possible to increase the reliability of defect inspection of the winding 1. In this way, by repeatedly storing and allocating inspection signals of the winding 1, it is possible to determine the defect location of the winding 1 continuously and with high reliability, even when the winding 1 is running at high speed.

[0042] FIG. 13 is a cross-sectional view showing an electric machine including a stator core, which is a wound member, and shows the state in which windings 1, which have been confirmed to be free of insulation defects using the above-described winding defect inspection device, have been wound around the stator core. FIG. 14 is a perspective view showing the state in which windings 1 are being wound around the teeth of a stator core 81 using a nozzle control unit 13. While FIG. 13 illustrates a rotating electric machine as an example of an electric machine 80, the electric machine 80 may also be a direct current motor. In this electric machine 80, the electric machine 80 is manufactured using a stator core 81 wound with windings 1, which have been confirmed to be free of insulation defects using the above-described winding defect inspection device. An electric machine can be manufactured by including such manufacturing steps in a manufacturing method for an electric machine.

[0043] As described above, the winding defect inspection device of this embodiment is applied to the winding process of a stator, which is the armature of a rotating electric machine or a direct current motor, as an example of an electric machine. Therefore, with the winding defect inspection device of this embodiment, even if the winding 1 runs at high speed, it is possible to continuously and reliably determine the defect locations in the entire winding 1 before winding, even if the winding 1 runs at high speed. Furthermore, it is possible to provide a stator, which is the armature of a rotating electric machine or a direct current motor, using a winding 1 with excellent reliability.

[0044] Embodiment 2. Figure 15 is a block diagram showing the configuration of a winding defect inspection device according to embodiment 2. In Figure 15, the same reference numerals as in Figure 1 indicate the same or corresponding parts. In the winding defect inspection device according to embodiment 2, sensor 2a is arranged above winding 1 and sensor 2b is arranged below winding 1, and sensors 2a and 2b are used to inspect defects around winding 1. Of the operations of the winding defect inspection device according to this embodiment, the operations of discharge detection unit 5, first through Nth computing units, and host machine 8 are the same as those described in embodiment 1.

[0045] The operation of the storage unit 6 will now be described. The storage unit 6 simultaneously receives detection signals from the sensors 2a and 2b via the discharge detection unit 5, stores these two detection signals separately, and assigns the data to the first to Nth computing units. An example of the assignment method in this case is to assign the inspection signal from the sensor 2a to the first half of the first to Nth computing units, and to assign the inspection signal from the sensor 2b to the second half of the first to Nth computing units. For example, if there are 10 computing units 7 (N=10), the first half will be the first to fifth computing units, and the second half will be the sixth to tenth computing units.

[0046] 15 is merely one example. That is, three or more sensors 2 may be arranged around the winding 1 at any angular interval. Furthermore, by arranging a plurality of defect detection units 9, including sensors 2a and 2b, the reliability of defect inspection of the winding 1 can be improved. In this way, by arranging a plurality of sensors 2 around the winding 1 and inspecting for defects around the winding 1, it is possible to continuously determine the defect locations in the winding 1, even when the winding 1 is running at high speed, and with even higher reliability than in the first embodiment.

[0047] Embodiment 3. Figure 16 is a block diagram showing the configuration of a winding defect inspection device according to embodiment 3. In Figure 16, the same reference numerals as in Figure 1 indicate the same or corresponding parts. In the winding defect inspection device according to embodiment 3, sensors 2a and 2b are arranged along the running direction of winding 1. Of the operations of the winding defect inspection device according to this embodiment, the operations of discharge detection unit 5, computing unit 7, and host machine 8 are the same as those described in embodiment 1.

[0048] 17 is a flowchart illustrating the operation of the winding defect inspection apparatus according to embodiment 3. A voltage is applied from power supply 3 to sensor 2a (first sensor) and sensor 2b (second sensor), and inspection of winding 1 for defects begins. First, in a discharge inspection and first sensor signal accumulation step (step S701), discharge detection unit 5 inspects for discharge between sensor 2a and winding 1, and accumulation unit 6 reads the output (first inspection signal) from discharge detection unit 5 and accumulates the inspection signal. When accumulation unit 6 has collected data for a certain length of time (e.g., 30 ms), the result is "YES" in an accumulation completion step (step S702), and the accumulation signal is passed to calculator 7 (step S703).

[0049] When the calculator 7 receives the inspection signal from the storage unit 6, it enters a calculation step by the calculator 7 (step S704) and begins analyzing the frequency characteristics of the inspection signal. At the same time, the storage unit 6 continues to store an inspection signal (second inspection signal) from the sensor 2b for the same duration (e.g., 30 ms) (step S705). When storage of the inspection signal from the sensor 2b is completed (step S706), the inspection signal is passed to the calculator 7 (step S707), and the calculator 7 enters the calculation step (step S704) again and begins analyzing the frequency characteristics of the inspection signal. At the same time, the storage unit 6 again begins storing the inspection signal from the sensor 2a.

[0050] Next, a method for determining the placement distance L between sensors 2a and 2b will be described. The placement distance L between sensors 2a and 2b is determined by the running speed of winding 1 and the time it takes for calculator 7 to receive the inspection signal and perform the calculation. If the running speed of winding 1 is V and the time it takes for calculator 7 to receive the inspection signal and perform the calculation, then placement distance L is calculated using the following equation (4). Here, a margin may be added to the time T it takes for calculator 7 to receive the inspection signal and perform the calculation.

[0051] L=V×T・・・・・・・・・・・・・・・(4)

[0052] For example, if the running speed of the winding 1 is V=500 mm / s and the sum of the reception time and calculation time of the calculator 7 is T=30 ms, the placement distance L is L=15 mm.

[0053] 16 is merely one example. That is, two or more sensors may be arranged along the running direction of the winding 1. Furthermore, by arranging a plurality of defect detection units 9, including sensors 2a and 2b, the reliability of defect inspection of the winding 1 can be improved. In this way, by arranging a plurality of sensors 2 along the running direction of the winding 1 and inspecting the winding 1 for defects, it is possible to continuously determine the defect locations in the winding 1, even when the winding 1 runs at high speed, and with even higher reliability than in the first embodiment.

[0054] Embodiment 4. Figure 18 is a block diagram showing the configuration of a winding defect inspection device according to embodiment 4. In Figure 18, the same reference numerals as in Figure 16 indicate the same or corresponding parts. The winding defect inspection device according to embodiment 4 has multiple sensors 2a-2d arranged above and below the winding 1, and inspects the periphery of the winding 1 for defects using the multiple sensors 2a-2d. Of the operations of the winding defect inspection device according to this embodiment, the operations of the discharge detection unit 5, computing unit 7, and host machine 8 are the same as those described in embodiment 3.

[0055] Next, the operation of the storage unit 6 will be described. While simultaneously receiving inspection signals from sensors 2a to 2d via the discharge detection unit 5, the storage unit 6 stores the inspection signals from sensors 2a and 2b in one storage area and the inspection signals from sensors 2c and 2d in another storage area, storing them separately and allocating them to the first and second computing units. As an example of an allocation method in this case, the inspection signals from sensors 2a and 2b can be allocated to the first computing unit, and the inspection signals from sensors 2c and 2d can be allocated to the second computing unit.

[0056] 18 is merely one example. That is, three or more sensors 2 can be arranged around the winding 1 at any angular interval. Furthermore, by arranging a plurality of defect detection units 9, including sensors 2a to 2d, the reliability of defect inspection of the winding 1 can be improved. In this way, by arranging a plurality of sensors 2a to 2d around the winding 1 and inspecting for defects around the winding 1, it is possible to continuously determine the defect locations in the winding 1, even when the winding 1 is running at high speed, and with even higher reliability than in the first embodiment.

[0057] Fifth Embodiment Figure 19 is a block diagram showing the configuration of a winding defect inspection device according to a fifth embodiment. In Figure 19, the same reference numerals as in Figure 1 indicate the same or corresponding parts. The winding defect inspection device according to the fifth embodiment has one sensor 2 arranged above the winding 1. One calculator 7a is arranged as the calculator 7. Of the operations of the winding defect inspection device according to this embodiment, the operations of the discharge detection unit 5, storage unit 6, and host machine 8 are the same as those described in the first embodiment.

[0058] The storage unit 6 receives the inspection signal from the sensor 2 via the discharge detection unit 5, stores the signal in a storage area, and then passes it on to the calculator 7a. This embodiment is applicable when the running speed of the winding 1 is slow, for example, 50 mm / s. When the speed of the winding 1 is slow, the time required for calculation in the calculator 7a is set to 30 ms, the same as in the first embodiment, and the accumulation time of the storage unit 6 is set to 30 ms. This means that the calculation by the calculator 7 will be completed by the time the accumulation of the detection signal in the storage unit 6 is completed, and therefore the winding 1 can be inspected for defects.

[0059] 19 is merely one example. A plurality of sensors 2 can be arranged around the winding 1. Furthermore, by arranging a plurality of defect detection units 9, the reliability of defect inspection of the winding 1 can be improved. In this way, by arranging sensors 2 around the winding 1, when the winding 1 is traveling at a low speed, the defect locations in the winding 1 can be determined continuously, accurately, and with high reliability.

[0060] Sixth Embodiment Figure 20 is a block diagram showing the configuration of a winding defect inspection device according to a sixth embodiment. In Figure 20, the same reference numerals as in Figure 1 indicate the same or corresponding parts. The winding defect inspection device according to the sixth embodiment has a voltage detection unit 44 that detects the voltage of the power supply 3, and instead of detecting the discharge current between the sensor 2 and the winding 1, it detects a change in the voltage of the power supply 3 when a discharge occurs between the sensor 2 and the winding 1, i.e., a fluctuation in the voltage applied between the sensor 2 and the winding 1, thereby inspecting for defects in the winding 1. Of the operations of the winding defect inspection device according to this embodiment, the operations of the discharge detection unit 5, the computing unit 7, and the host machine 8 are the same as those described in the first embodiment.

[0061] Next, the principle of the method for inspecting the winding 1 for defects by detecting voltage changes in the power supply 3 will be described. Consider the case where an AC voltage, for example, an AC voltage with a frequency of 60 Hz, is applied to the winding 1 from the power supply 3. FIG. 21A shows the change in power supply voltage when there is no defect in the winding, and FIG. 21B shows the change in power supply voltage when there is a defect in the winding. The voltage of the power supply 3 is detected by resistors 45 and 46 of the voltage detection unit 44, and noise processing and other processes are performed in the discharge detection unit 5, and the processed signal is passed to the storage unit 6. When there is no defect in the winding 1, the voltage of the power supply 3 does not change, and therefore a normal voltage signal 47 shown in FIG. 21A is detected. When the normal voltage signal 47 is processed using the wavelet transform of (Equation 2) described in embodiment 1, only a frequency of 60 Hz is detected. On the other hand, when there is a defect in the winding 1, a discharge is likely to occur near the peak of the voltage of the power supply 3. When this occurs, the voltage output terminal and the ground terminal (GND) of the power supply 3 are shorted, causing the output voltage of the power supply 3 to change suddenly, resulting in the detection of a voltage signal like the defect voltage signal 48 shown in FIG. 21B. If the frequency of the voltage signal between times t1 and t2 and between times t3 and t4 is high, for example, 10 MHz, and the defect voltage signal 48 is processed using the wavelet transform of (Equation 2) described in the first embodiment, the 10 MHz frequency as well as the 60 Hz frequency will be detected, thereby determining that there is a defect in the winding 1. Thus, by detecting a voltage drop in the power supply 3, which is easy to detect, the location of a defect in the winding 1 can be detected continuously and accurately, even when the winding 1 is running at high speed, and can be determined with even higher reliability. Furthermore, a stator, which is an armature for a rotating electric machine or a direct current motor, can be provided that uses a winding 1 with excellent reliability.

[0062] Although various exemplary embodiments and examples are described in this application, the various features, aspects, and functions described in one or more embodiments are not limited to the application of a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless variations not illustrated are contemplated within the scope of the technology disclosed in this specification. For example, this includes cases where at least one component is modified, added, or omitted, or where at least one component is extracted and combined with components of another embodiment.

[0063] 1 Winding, 2 Sensor, 3 Power supply, 4 Discharge current or charge amount, 5 Discharge detection unit, 6 Storage unit, 7 Calculator, 11 Wound member, 17 Defective part, 80 Electric machine, 81 Stator core.

Claims

1. A defect inspection device for a winding used in manufacturing a winding component by winding the winding around a wound member, the defect inspection device comprising: a sensor located midway along a path from one end of the winding to the wound member; a discharge detection unit that detects a discharge current, charge amount, or fluctuations in the voltage generated by a voltage applied between the sensor and a defective portion of the winding; a storage unit that stores an inspection signal output from the discharge detection unit; and a calculator that allocates the inspection signal stored in the storage unit and calculates the time required to identify the defect location in the winding by analyzing the frequency characteristics of the inspection signal.

2. The winding defect inspection device according to claim 1, comprising a plurality of said computing units.

3. A winding defect inspection device according to claim 1 or 2, wherein the computing unit determines the defect position of the winding based on the time for identifying the defect position and the running speed of the winding.

4. A winding defect inspection device according to any one of claims 1 to 3, wherein a plurality of the sensors are arranged around the winding, and a plurality of the inspection signals are detected by the discharge detection unit.

5. A winding defect inspection device according to any one of claims 1 to 4, wherein a plurality of the sensors are arranged along the running direction of the winding, the storage unit stores a first inspection signal from a first sensor, and passes the stored first inspection signal to the computing unit, causing the computing unit to perform a calculation, and after the storage unit passes the first inspection signal to the computing unit, the storage unit stores a second inspection signal from a second sensor, and passes the stored second inspection signal to the computing unit, causing the computing unit to perform a calculation.

6. A winding defect inspection device as described in any one of claims 1 to 5, wherein the computing unit determines whether defect positions of the winding are adjacent when the winding is wound around the wound member based on the running speed of the winding and the vertical and horizontal widths of the wound member.

7. A method for manufacturing an electric machine, in which the windings that have been confirmed to be free of insulation defects using the winding defect inspection device described in any one of claims 1 to 6 are wound around a stator core, thereby manufacturing an electric machine equipped with the stator core.

8. A winding defect inspection method for inspecting defects in a winding when manufacturing a winding component by winding the winding around the wound member, comprising: a sensor arranged midway along a path from one end of the winding to a wound member; a discharge detection unit that detects a discharge current, charge amount, or fluctuations in the voltage generated by a voltage applied between the sensor and a defective portion of the winding; a storage unit that stores an inspection signal that is an output from the discharge detection unit; and a plurality of computing units that perform calculations based on the inspection signals stored in the storage unit, a winding defect inspection method in which the storage unit stores the inspection signal over a fixed time period and assigns the stored inspection signal to a first computing unit, causing the first computing unit to start an operation to analyze the frequency characteristics of the inspection signal; simultaneously, the storage unit starts storing the inspection signal over a second fixed time period; when the storage unit has completed storing the second inspection signal, it assigns the stored inspection signal to a second computing unit, causing the second computing unit to start an operation to analyze the frequency characteristics of the inspection signal; the above operation is repeated the number of times equal to the number of computing units, and after the inspection signal has been assigned to the last computing unit, the assignment of the inspection signal is switched to the first computing unit; and the plurality of computing units then analyze the frequency characteristics of the inspection signal, thereby calculating the time required to identify a defect location in the winding.

Images