Lathe system diagnostic device, power conversion device, lathe system and lathe system diagnostic procedure

Abstract

Diagnostic device (4) for diagnosing a condition of a lathe system, wherein the lathe system comprises a lathe (3), a cable, a power conversion device (7) and a power source (1) which are electrically connected to each other and The power conversion device contains the following: a current sensor (10a, 10b) configured to measure a current of the power conversion device; and a control unit configured to set a switching state of the power conversion device based on an output from the current sensor; and the diagnostic device comprises a diagnostic unit (6) for diagnosing a condition of the lathe system based on an output from the current sensor, and wherein The diagnostic unit is configured to diagnose the condition of the lathe system using output information from the current sensor and information acquired by the control unit indicating a time elapsed since a switching point, which serves as a reference to any specified phase of the power conversion device.

Description

[Technical field]

[0001] The present invention relates to a lathe system diagnostic device, a power conversion device equipped with a diagnostic functionality, a lathe system and a lathe system diagnostic method. [Background of the invention]

[0002] If an unexpected fault occurs in a lathe, such as in a motor (electric motor) or generator installed in a production plant, unplanned repairs or replacements of the lathe will be necessary, and the operating rate of the production plant will need to be reduced or the production plan modified. Similarly, if a power conversion device or a cable connected to the lathe fails, unplanned repairs or replacements will also be necessary, and the operating rate of the production plant will need to be reduced or the production plan modified.

[0003] To prevent unexpected failures of a lathe system (lathes and their accessories (cables, power conversion devices)), lathe systems are appropriately stopped and diagnosed offline (offline diagnostics) to assess the degree of wear and prevent unexpected failures to a certain extent. However, stopping the lathe system reduces the operating rate of the production plant. Additionally, depending on the type of wear, it may only become apparent when a voltage is applied, making offline diagnostics difficult. Therefore, there is a need to diagnose the condition of a lathe system during operation.

[0004] Japanese patent publication JP 2011 - 229 322 A (Patent document 1) discloses a diagnosis based on current information for a lathe system and discloses a technique for estimating that a cable in a motor system is short-circuited when a sum of values ​​from current sensors attached at two points exceeds a certain threshold.

[0005] Furthermore, Japanese patent application JP 2015-56918A (Patent Document 2) describes a vehicle drive unit comprising a rotating electronic machine with a rotating shaft connected to a power transmission path of the vehicle and including a plurality of phase coils and a power conversion device with a plurality of phase arms. The rotating electronic machine also has at least one current sensor located in a phase between a phase coil and a phase arm to detect and output the phase current, and a control unit for controlling the power converter based on the information obtained from the current sensor.

[0006] Furthermore, Japanese patent application JP 2009-50059A (Patent Document 3) discloses a fault detector designed to detect faults in a drive control system for an AC lathe driven by a power converter. This fault detector comprises current sensing means arranged in the power converter's connecting means, a switching command generation means that outputs a switching command signal to the power converter's switching elements to apply voltage to all three phases of the AC lathe, and generates a special switching command signal in which two of the three-phase voltages have the same polarity and the remaining voltage is zero, as well as a fault detection means that, based on the currents detected by the current sensing means, identifies a fault when the special switching command signal is output.

[0007] Furthermore, German patent application DE 10 2014 107 094 A1 (patent document 4) describes an electrical machine that is electrically connected to a rectifier / inverter via a multiphase power circuit. Based on this, a method for monitoring the multiphase power circuit is also motivated, which non-intrusively adjusts a commanded alternating current from the rectifier / inverter after a prescribed period of time and compares a measured value of the alternating current in the multiphase power circuit with a minimum threshold value. [List of prior art][Patent literature] [Patent Document 1] Japanese Patent Publication JP 2011 - 229 322 A [Patent document 2] Japanese patent disclosure JP 2015 - 56 918 A [Patent document 3] Japanese patent disclosure JP 2009 - 50 059 A [Patent document 4] German patent disclosure DE 10 2014 107 094 A1 [Summary of the invention][Technical problem]

[0008] The technology disclosed in patent document 1 requires the installation of an additional current sensor. Since the diagnosis is based on the sum of two current sensor readings, it is not possible to perform additional diagnostics on wear mechanisms other than cable short circuits. Furthermore, a further improvement in diagnostic accuracy is desired.

[0009] Accordingly, it is an object of the present invention to provide a diagnostic device and a diagnostic method for a lathe system, a power conversion device equipped with a diagnostic functionality, and a lathe system that solves the aforementioned problems. [Solution to the problem]

[0010] To solve the problem of the present invention, the contents according to the independent claims are proposed. Dependent claims also correspond to preferred embodiments of the invention. [Advantageous effects of the invention]

[0011] According to the present invention, the addition of a current sensor is unnecessary and the condition of the lathe system can be diagnosed. [Brief description of the drawings] [ Fig. 1] Fig. Figure 1 is a diagram representing a configuration of embodiment 1. [ Fig. 2] Fig. Figure 2 is a diagram that represents an evaluation procedure of embodiment 1. [ Fig. 3] Fig. Figure 3 is a diagram representing a diagnostic result of embodiment 1. [ Fig. 4] Fig. Figure 4 is a diagram that represents a configuration of embodiment 2. [ Fig. 5] Fig. Figure 5 is a schematic diagram of a measurement method of embodiment 2. [ Fig. 6] Fig. Figure 6 is a diagram that represents a configuration of embodiment 2. [ Fig. 7] Fig. Figure 7 is a diagram showing Lissajous figures for normal and worn states of embodiment 3. [ Fig. 8] Fig. Figure 8 is a schematic diagram of an analysis method of embodiment 3. [ Fig. 9] Fig. Figure 9 is a diagram showing Lissajous figures for normal and worn states of embodiment 3. [ Fig. 10] Fig. Figure 10 is a diagram showing frequency spectra for normal and worn states of embodiment 3. [ Fig. 11] Fig. Figure 11 is a diagram showing Lissajous figures for normal and worn states of embodiment 4. [ Fig. 12] Fig. Figure 12 is a schematic diagram of a measurement method of embodiment 5. [ Fig. 13] Fig. Figure 13 is a schematic diagram of an analysis procedure of embodiment 5. [ Fig. 14] Fig. Figure 14 is a configuration diagram of embodiment 6. [ Fig. 15] Fig. Figure 15 is a configuration diagram of embodiment 7. [ Fig. 16] Fig. Figure 16 is a configuration diagram of embodiment 8. [Description of the embodiment(s)]

[0012] Since the failure rate of lathes, such as motors, is high due to insulation wear, an additional current sensor is conventionally provided to diagnose the condition and perform desired measurements. However, although power conversion devices such as inverters include a control current sensor, the problem arises that it is difficult to detect phase currents such as weak leakage currents and the like, making it difficult to use it for diagnostic purposes. Accordingly, the present inventors considered a current component that can detect signs of wear in a lathe system and also took into account that by obtaining the current from the control current sensor of the power conversion device, an additional current sensor becomes unnecessary.

[0013] In the lathe system, a lathe, a power conversion device, and a power source are electrically connected by a cable, and the power conversion device includes a current sensor configured to measure a current in a connection path between the lathe and another device; and a control unit configured to set a switching state based on an output from the current sensor.

[0014] Therefore, the control current sensor transmits the transmitted information to the diagnostic device, and this current information is used to diagnose the lathe system. In other words, the lathe system's diagnostic device includes a diagnostic unit that diagnoses the system's condition based on the output of the control current sensor used in the power conversion device. As a result, this eliminates the need to install an additional current sensor and also avoids the need to modify the existing lathe system. The diagnostic device can be equipped with a data extraction unit or a data storage unit in addition to the diagnostic unit for diagnosing the lathe based on the current sensor's output information.

[0015] The diagnostic device and the power conversion device can be separate units; however, if the inverter microcomputer has excess power, all or part of the diagnostic device's functions can be integrated into the inverter control microcomputer. As a result, and with a saving of space, the control signal can be easily acquired.

[0016] A diagnosis based on the acquired current waveform information is performed by, for example, Fourier-transforming a waveform of one period or more, extracting a change in amplitude at a given frequency, and calculating the magnitude or rate of change of the amplitude. If a fine signal is required to observe signs of wear in the lathe system, it is possible to seemingly increase the measurement frequency by using the periodicity of the current sensor output and superimposing several periods without increasing the sampling rate. By acquiring current values ​​for several different elapsed times from the reference point over multiple periods, a measurement can be performed that is equivalent to a measurement with an increased sampling rate.

[0017] The reference point can be determined from the current sensor output or based on switching information received from the control unit. If the current waveform exhibits periodicity, data for 1 / N periods, such as 1 / 4 or 1 / 2 periods, from the reference point in the current waveform can be used for diagnostic purposes. In addition to a single phase of the lathe's current, two-phase and three-phase current values ​​can also be used for diagnostic purposes. It is also possible to perform diagnostics using a combination of these phase currents. To further improve diagnostic accuracy, signals from the temperature and humidity sensors can be acquired and fed into the diagnostic device.

[0018] If a current sensor range to be used for control purposes is defined, it is preferable to perform diagnostics based on an output that is not used for control. For example, in some cases, the high-frequency components (overshoot waveforms) caused and generated by switching the control system can be excluded from the lathe's control data. Diagnostics based on the current sensor output in the overshoot waveform section allow the current sensor output to be used without affecting the control information.

[0019] Details of configurations for carrying out the present invention (hereinafter referred to as "embodiments") are described below with reference to the figures, as applicable. Furthermore, the following description merely provides examples of the embodiments, and the scope of the present invention is not limited to the embodiments described below. EXECUTION FORM 1

[0020] Fig. 1 is a configuration diagram of embodiment 1. In the lathe system of Fig. In the following diagram, a power source 1, a cable 2, a power converter 7, and a lathe 3 are electrically connected, and a three-phase AC voltage is output at the power converter 7. The output of the three-phase AC voltage is controlled by adjusting the timing at which the switching element of the inverter is actuated, such that the speed and torque of the motor are the desired values. The control is determined based on the control information and the information about the current output by the inverter, and a control current sensor 10 is positioned to measure the value of the current flowing in any selected phase in order to acquire the current information of the motor.

[0021] In the present embodiment, the diagnostic device 4 is connected to the lathe system, and the output of the control current sensor 10 is split and used to diagnose the lathe system. The current information received by the diagnostic device can be an analog value from the control current sensor 10a or a value digitized by an analog-to-digital converter for use in the control system. For example, the analog value can include a voltage waveform output and a current waveform output from the control current sensor 10a. Additionally, a waveform obtained by splitting a voltage waveform or a current waveform according to the design of the diagnostic device can be input into the diagnostic device.Additionally, in cases where an analog value is used, it is preferred that the input impedance of the diagnostic device 4 is sufficiently increased with respect to the input impedance of the control unit, or that the signals acquired by the diagnostic device and the control unit are appropriately corrected according to the input impedance of the diagnostic device and the control unit.

[0022] The diagnostic device 4 according to the present embodiment includes a diagnostic unit 6, and the diagnostic unit 6 performs a diagnosis based on the current waveform information measured by the control current sensor 10a. By using the control current sensor 10a also as a diagnostic sensor, the installation of an additional current sensor becomes unnecessary, and the diagnostic function can be provided in a cost-effective manner. Since the diagnosis uses the phase current value as a basis, it is also possible to diagnose the condition (such as aging wear, bearing wear, and imbalance) of the lathe system with high accuracy.

[0023] The following diagnostic procedure, which constitutes the diagnostic procedure of the present embodiment, involves Fourier-transforming a waveform of one period or more based on current waveform information, observing the change in the amplitude of a frequency of interest, and detecting wear based on the magnitude / rate of change. The frequency of interest can be a single frequency or multiple frequencies. An example of focusing on a single frequency or multiple frequencies is, for instance, diagnostics using motor signature analysis (MCSA). MCSA, also known as current system analysis, is a technique for detecting signs of wear, such as bearing wear and imbalance, based on changes over time in the frequency spectral ranges of current waveforms.

[0024] In the diagnostic unit, the waveform data of the control current sensor 10a is Fourier-transformed, and signs of wear are detected based on the increase or decrease in the amplitude or frequency of interest. Additionally, if the characteristic frequency is unknown, a frequency that is statistically significantly different from the Fourier transform spectrum of the control current sensor 10a is sought, and advanced wear is determined if the statistical degree of separation exceeds a predefined threshold.For example, a procedure can be considered in which the difference between the center of the spectrum amplitude distribution for a given frequency, learned as "normal," and the spectrum amplitude of the frequency to be diagnosed is divided by the spectrum amplitude distribution for a given frequency learned as normal. Then, if the resulting value exceeds a predetermined threshold, it can be determined that wear is advanced. It should be noted that, although an example has been described in which the MCSA appears as a peak value at a given frequency, in the case of delocalized wear, it need not appear as a peak value but may, in some cases, appear as a rise around the driving frequency.

[0025] Insulation wear in a lathe system (a lathe and its accessories (cables, power converters)) can, in some cases, exhibit a peak at a specific frequency, but it primarily affects the current in the high-frequency range as a change in the impedance of the current path. Additionally, if the resonant frequency of the path through which the current flows changes due to the impedance change, and the current contains many frequency components, the change can occur at a specific frequency. In particular, in the case of insulation wear, signs of wear tend to appear in components that operate at high frequencies.

[0026] Accordingly, to detect insulation wear in lathe systems, the waveforms obtained in the diagnostic device 4 can be Fourier-transformed and compared based on the change in the signal amplitude of the high-frequency components of the obtained signal. The frequency range of interest is not particularly limited; however, due to the signal-to-noise ratio, it is preferred that the frequency be as high as possible, in particular a few kHz or more. For example, the carrier frequency component of the inverter and the frequency component of the overshoot waveform (approximately a few MHz) at the time of inverter switching can be taken into account.

[0027] The effects of the present embodiment are described below based on current waveforms measured before and after insulation wear on a lathe in which the insulation wear was actually advanced. The following describes the insulation wear of the lathe; however, the same changes have been confirmed in cases where insulation wear occurs on the cable and in cases where insulation wear occurs on the power conversion device.

[0028] First, a signal from a control current sensor is input into the diagnostic device 4, and the power conversion device 7 is connected to each of two types of motors: a normal motor and a motor that has undergone insulation wear. To learn the normal state, a current waveform corresponding to several periods of the fundamental frequency was initially recorded while the normal motor was connected. The resulting current waveform was then divided into 100 files, one for each of the 5 periods of the fundamental frequency, and a Fourier transform was performed on each file to extract the amplitude of the arbitrarily selected frequency components.

[0029] In this measurement, the inverter was set such that the fundamental frequency was 50 Hz. Ideally, the frequency component to be extracted should be as high as possible due to the signal-to-noise ratio, and in particular, several kHz or more are desirable. In addition to 2 kHz, 4 kHz, 6 kHz, 8 kHz, and 10 kHz, which are integer multiples of the inverter's carrier frequency of 2 kHz, integer multiples of the fundamental frequency with a relatively low frequency but a high signal-to-noise ratio were selected, such as 50 Hz, 500 Hz, and 1 kHz.

[0030] Subsequently, a histogram of the amplitudes was generated, which was in Fig. As shown in Figure 2, a probability density function (here a normal distribution) was created for the selected frequency components and fitted to the distribution of the histogram to obtain the variance and mean for each arbitrarily chosen frequency. The variance and mean obtained for learning the normal state are subsequently referred to as the "learned normal distribution mean" and the "learned normal distribution variance," respectively. Then, the waveform of the normal state was re-acquired for five periods of the fundamental wave, a Fourier transform was performed, and the amplitude of the arbitrarily chosen frequency components was extracted. The amplitude was then transformed according to equation (1). [Formula 1] Amplitude – learned normal distribution mean – learned normal distribution variance

[0031] Similarly, for motors with insulation wear, current waveforms were recorded for five periods of the fundamental frequency, a Fourier transform was performed, and the amplitude of arbitrarily chosen frequency components was extracted. The obtained values ​​are plotted on the vertical axis, and the frequency is plotted on the horizontal axis of the graph. Fig. 3. These measurement results confirmed that signs of insulation wear appeared significant at the carrier frequency of 2 kHz.

[0032] As described above, the configuration of the present embodiment makes it possible to diagnose insulation wear on lathes. It should be noted that, depending on the measuring system, if a device emitting noise with a carrier frequency of 2 kHz (e.g., an inverter with a carrier frequency of 2 kHz) is operating nearby, the 2 kHz noise in the signal detected by the control current sensor 10a increases, and the learned normal distribution variance also increases. In such cases, signs of insulation wear at 2 kHz are less likely to appear. Therefore, it is desirable to diagnose multiple frequency components and evaluate them based on their variance. EXECUTION FORM 2

[0033] In embodiment 1 and Fig. In Figure 1, an example was described in which the diagnostic device 4 is an independent device separate from the power conversion device 7. However, in embodiment 2, the diagnostic device 4 and the control unit 8, which are composed of the extraction unit 5 and the diagnostic unit 6, differ from embodiment 1 in that they are formed in the power conversion device 7 in a jointly used microcomputer, as shown in Figure 1. Fig. Figure 4 illustrates this. By combining the diagnostic device and the control unit into a single microcomputer, costs can be reduced and space can be saved compared to cases where the diagnostic device and the control unit are separate.

[0034] Furthermore, it is also possible to synchronize the signals between the diagnostic device and the control unit, which makes it possible to reduce the sampling rate required by the diagnostic device. The principle by which the sampling rate can be kept low is described below.

[0035] The diagnostic device 8 adjusts the switching point so that the motor's rotational speed and torque reach the desired values ​​and can determine from the switching instruction whether the switching point settles into a stable state. If a stable state can be determined, the current value of each period, detected by the control current sensor 10a, does not fluctuate significantly, and the lathe possesses sufficient accuracy for use in diagnosing the lathe system.

[0036] The diagnostic device 4 can be equipped with a filter 11, such as a bandpass filter, for extracting a specific frequency according to the frequency of interest. For example, wear of the lathe bearing affects the low-frequency components, and wear of the insulation affects the high-frequency component, resulting in a corresponding signal. By providing a filter for extracting such a characteristic frequency, it is possible to accurately detect a target condition or signs of wear. If the filter 11 is provided, it is preferred to place it between the data extraction unit 5 and the sensor.

[0037] Fig. Figure 5 is a schematic diagram of current data acquired by the extraction unit of the present embodiment. In a situation where a stable state can be determined, the extraction unit 5 acquires current values ​​for time points elapsed since the switching time, which serves as a reference for the control device with Δt1 as a first time, Δt2 as a second time, Δt3 as a third time, and Δtn as an nth time. Fig. 5(a) to Fig. 5(c)). By measuring information for the current waveform several times after a time arbitrarily changed from a reference time and reconstructing it into time series data ( Fig. 5(d)) Data can be obtained that are equivalent to the case of measuring Δt1, Δt2, Δt3,... Δtn at a high frequency in a single measurement. That is, it is possible to perform a measurement with an apparently increased sampling rate using information about the phase elapsed since the switching time, which serves as a reference for the arbitrarily determined phase.

[0038] Since the microcomputer 9 of the power conversion device 8 must perform not only data measurement but also control processing, it is impractical to always perform a current measurement. Accordingly, by changing the point in time at which the measurement processing is carried out in each period, it is possible to perform the measurement with an apparently increased sampling rate. It should be noted that the same effect can be obtained even if such measurement processing is carried out in a case where the diagnostic device is provided separately from the power conversion device.

[0039] Regarding the method for analyzing the waveform of the obtained control current sensor 10a ( Fig. 5(d)) it is possible to use a procedure such as MCSA, where a Fourier transform is used to detect signs of wear from increases or decreases in the amplitude of the frequency of interest, or a statistical procedure can be used in which it is determined that wear is progressing if a value obtained by dividing the difference between the center of the amplitude variance of the spectrum of a frequency learned as normal and the amplitude of the spectrum of the frequency to be diagnosed by the amplitude variance of the spectrum of a given frequency learned as normal exceeds a predetermined threshold. EXECUTION FORM 3

[0040] In embodiments 1 and 2, a condition of the lathe system, such as a malfunction, is analyzed by a control current sensor 10a. However, embodiment 3 differs in that two control current sensors 10a and 10b, installed with different phases, are used, and a diagnosis is performed using these two signals. It should be noted that, although Fig. 6 is an example in which the diagnostic unit 6 is configured in the same microcomputer as the control unit 8, the diagnostic unit 6 can be configured independently of the microcomputer of the control unit, or it can be provided as a diagnostic device separate from the power conversion device.

[0041] As in embodiments 1 and 2, current information is acquired by the control current sensors 10a and 10b and the presence or absence of wear of the lathe system can be detected by analyzing its waveforms.

[0042] Each of the frequency amplitudes of the control current sensors 10a and 10b can be analyzed separately and it can be determined that wear is present if the amplitude of at least one of the control current sensors exceeds a threshold value, or, to avoid false alarms, it can be determined that wear is present if the average value of the amplitude of a certain frequency component of the control current sensors 10a and 10b exceeds a certain threshold value.

[0043] The procedure for analyzing each of the waveforms is the same as in embodiments 1 and 2, and it is possible to use a method such as MCSA, where a Fourier transform is used to detect signs of wear from increases or decreases in the amplitude of the frequency of interest, or a statistical procedure can be used in which it is determined that wear is progressing if a value obtained by dividing the difference between the center of the amplitude variance of the spectrum of a frequency learned as normal and the amplitude of the spectrum of the frequency to be diagnosed by the amplitude variance of the spectrum of a given frequency learned as normal exceeds a predetermined threshold.

[0044] Additionally, a diagnostic procedure focused on changes in Lissajous figures can be mentioned as a method in which two current sensor information pieces are used and no frequency analysis is performed. Fig. Figure 7 represents the Lissajous curves of a normal state and a state of insulation wear. As insulation wear progresses, the value of the current flowing at the switching time of the power converter 7, i.e., the high-frequency component of the current, increases, and the Lissajous figure changes. By visualizing these changes as a graph, the diagnostic unit can inform the user of the presence or absence of an anomaly, i.e., the presence or absence of a change in the Lissajous figure. In addition to visualizing the Lissajous figure itself, a method can also be used to evaluate the area of ​​the region enclosed by the Lissajous figure, the area of ​​the Lissajous figure, and the extension of the Lissajous curve in an arbitrarily defined region.

[0045] Additionally, to detect tiny changes and to check for the presence or absence of the initial stages of wear described above, the application of machine learning can be effective, and although the algorithms for machine learning are not particularly limited, for example a method known as the local subspace method can be applied.

[0046] In this procedure, as it is described in Fig. As shown in Figure 8, data to be defined as a normal state are prepared, two points near the data to be diagnosed are extracted, and the length of a line obtained from the data to be diagnosed by subtracting a straight line perpendicular to the line connecting these two points is defined as the degree of anomaly. The data to be diagnosed represent an instantaneous value of the control current sensors 10a and 10b at a specific time. A case in which the degree of anomaly of the data to be diagnosed exceeds a certain value, or the average value of the degree of anomaly over an arbitrarily defined period of the fundamental wave, can be defined.

[0047] The average values ​​of the degrees of anomaly of five periods of the Lissajous figure in the normal state and in the insulation wear state, which in Fig. The values ​​shown in 7 are 0.13 and 0.48, respectively. For example, if the threshold is set to 0.3 and this threshold is exceeded, it can be determined that an anomaly exists.

[0048] Similar diagnoses are also possible for bearing wear. Fig. Figure 9 represents a Lissajous figure for both the normal and initial states of engine bearing wear. It was ensured that 11 Lissajous figures had changed due to wear. Furthermore, as in Fig. Figure 10 shows that a peak value appears near the fundamental frequency in the frequency spectrum.

[0049] In Fig. 7 and Fig. 9. Although motor wear has been explained as an example, when the impedance of the lathe system changes, the value of the flowing current changes, and this appears as a Lissajous figure or as a frequency spectrum. As a result, the diagnosis based on the output of the control current sensor of this embodiment can also diagnose wear of the motor system.

[0050] Additionally, if a mechanical loss increases due to grease wear, bearing damage, or the like, the electrical output, i.e., the output of the control current sensor, changes accordingly. Therefore, according to the present embodiment, it is possible to detect even anomalies resulting from grease wear, bearing damage, or the like.

[0051] When a diagnostic test is performed, the signal acquired by the diagnostic device changes sensitively depending on the environment, e.g., temperature and humidity. To increase diagnostic accuracy, it is desirable to compare the environment during normal conditions with the environment during the measurement of the object being diagnosed. Specifically, methods include using a temperature sensor or a humidity sensor, or acquiring data from existing sensors at multiple times when temperature and humidity differ, and then performing a measurement under similar environmental conditions.If it is difficult to prepare data from multiple times with different temperatures and humidity levels, it is possible to suppress the occurrence of false alarms by confirming whether the change in the current signal detected by the diagnostic device is significantly different from the change in temperature and humidity. EXECUTION FORM 4

[0052] In embodiments 1 to 3, a diagnosis was performed using the entire duration of the current waveform. Since the phase of the waveform can be predicted from the time of the control device switching, only a segment of the fundamental period can be used for diagnosis. In embodiment 4, a diagnosis of a lathe using a segment of a fundamental period is described. The amount of data used to learn the normal state can be reduced by using only a segment of the period (e.g., a quarter of the period, a half of the period, a third of the period, etc.).

[0053] For example, Fig. 11 data points represent a 1 / 4 period. In this way, only data for a 1 / 4 period need to be collected, extracted, and used for diagnosis. In this case, it is desirable to adjust the normal state data so that they are at least slightly larger than a 1 / 4 period. This is because if the normal state data are in the same range as the diagnostic data, there may be cases where it can be difficult to make comparisons with the normal state or to diagnose the condition in the borderline section of the diagnostic data. For example, if the diagnosis is performed using normal data close to the diagnostic data as a reference, there is a possibility that the nearest normal value lies in a range outside the end range of the diagnostic data. Accordingly, it is preferable to use normal data that include these values.

[0054] In this embodiment, by using only a portion of the periodic diagnostic data, the amount of data used to learn the normal state can be reduced. EXECUTION FORM 5

[0055] In embodiments 1 to 3, the diagnosis is performed using the entire duration of the current waveform, while in embodiment 4 only a portion of the fundamental waveform is used. However, the diagnosis can also be performed using the overshoot waveform immediately after switching to a specific phase. Additionally, if the change in the overshoot waveform immediately after switching is not strongly phase-dependent, the diagnosis can be performed using an overshoot waveform of a different phase. Embodiment 5 describes a diagnosis of a lathe using an overshoot waveform immediately after switching.

[0056] As in the further embodiments described above, in cases where the diagnosis is performed using the entire period of the current waveform, the control unit 8 of the power conversion device 7 acquires data in such a way that the overshoot waveform of the control current sensor 10 is avoided. In contrast, in embodiment 5, the current data is acquired with respect to the switching point at which the overshoot waveform is generated.

[0057] In particular, since it has been determined that information regarding insulation wear is contained in the high-frequency component of the overshoot, it is possible to perform a highly accurate condition diagnosis of the lathe's insulation. Furthermore, by prioritizing the acquisition of overshoot wave data, it is possible to carry out the necessary diagnosis while keeping the amount of data that needs to be acquired to a minimum.

[0058] Fig. Figure 12 is a schematic diagram of current data acquired by the acquisition unit of the present embodiment. In a situation where a stable state can be determined, the extraction unit acquires 5 current values ​​for time points elapsed since the switching time, which serves as a reference for the control device with Δt1 as a first time, Δt2 as a second time, Δt3 as a third time, and Δtn as an nth time. Fig. 12(a) to Fig. 12(c)). By performing a measurement of the information from overshoot waveforms generated at the time of switching for each of several times after any period of time has elapsed since the switching time, the information can be converted into time series data ( Fig. 12(d)) can be reconstructed and it is possible to perform a measurement with an apparently increased sampling rate.

[0059] It should be noted that the number of control current sensors used for diagnosis is not particularly limited; however, if only one control current sensor is present, the diagnosis uses the change in the overshoot waveform as a basis, and if two or more control current sensors are present, the diagnosis can use the change in the overshoot waveform of these phases as a basis.

[0060] If necessary, the change in the waveform can be diagnosed by a value parameterized as a characteristic set, or machine learning can be applied using the characteristic set as an input parameter.

[0061] Examples of the parameterized characteristic set include the frequency of an overshoot waveform and the time constant of the damping. Additionally, when the outputs of multiple control current sensors are used for diagnosis, the change in the Lissajous figures of a section of the period can be diagnosed using a method such as the local subspace method. Fig. Figure 13 presents an example of the local subspace method using the output of a three-phase current sensor. In this method, data to be defined as a normal state are prepared, three points near the data to be diagnosed are extracted, the length of a line obtained from the data to be diagnosed by subtracting a straight line perpendicular to the plane connecting these three points is defined as the degree of anomaly, and a diagnosis is performed. EXECUTION FORM 6

[0062] Embodiment 6 differs from embodiment 5 in that, as in Fig. As shown in Figure 14, the extraction unit and the control unit 8 of the diagnostic device are mounted in the microcomputer 9, and the diagnostic unit 6 is configured outside the microcomputer 9. By transmitting the measurement data acquired by the microcomputer 9 to a diagnostic unit located outside the microcomputer, it is possible to reduce the processing power of the microcomputer. In addition to the instantaneous value of the current, data indicating how much of the phase has elapsed since the reference switching is transmitted to the diagnostic unit 6, allowing the waveform required for frequency analysis to be reconstructed. EXECUTION FORM 7

[0063] In embodiment 7, as in Fig. As shown in Figure 15, the diagnostic device 4 is not configured in the same microcomputer as the control device 8, but is configured separately. By providing the power conversion device 7 and the diagnostic device 4 separately and sending the switching information to 5, the processing power of the microcomputer and the sampling rate of the diagnostic device 4 can be reduced.

[0064] Information corresponding to the data indicating how many phases have elapsed since the switching time of the control device, and the signal from the control current sensor 10a, are input from the control device 8 to the diagnostic device 4. The current waveform can be reconstructed from the signal input to the diagnostic device 4, and the condition of the lathe system can be diagnosed.

[0065] If the current signals acquired by the control unit 4 are rearranged in phase starting from the reference switching time, frequency analysis can be applied, since the phase intervals are intended to be equal. Conversely, in the case of a discontinuity, frequency analysis is applied after interpolation between data has been performed to create equal intervals. If data with the same phase as the data to be compared has been obtained, the waveforms can be directly compared. EXECUTION FORM 8

[0066] In embodiment 8, an example of a diagnostic device 4 is described which, in addition to the extraction unit 5 and the diagnostic unit 6, has a data storage unit 12. As in Fig.As shown in Figure 16, the data storage unit 12 is provided in the diagnostic device 4. Once the data acquired by the extraction unit 5 has been stored in the data storage unit 12, the data is transmitted to the diagnostic unit 6. According to this configuration, the diagnosed data is accumulated in the data storage unit in a time series, and a comparison with previous normal data, such as the trend of time-dependent changes, becomes easier. Furthermore, by configuring the data storage unit 12 to store data with non-volatile memory, it becomes unnecessary to continuously supply power to the diagnostic device 4. [List of reference symbols] 1 power source 2. Cable 3 lathe 4 Diagnostic device 5 extraction units 6 Diagnostic Unit 7 Power conversion device 8 Control unit 9 microcomputers 10a, 10b Control current sensor 11 filters 12 Data storage unit

Claims

[1] Diagnostic device (4) for diagnosing a condition of a lathe system, wherein the lathe system comprises a lathe (3), a cable, a power conversion device (7) and a power source (1) which are electrically connected to each other and The power conversion device contains the following: a current sensor (10a, 10b) configured to measure a current of the power conversion device; and a control unit configured to set a switching state of the power conversion device based on an output from the current sensor; and the diagnostic device comprises a diagnostic unit (6) for diagnosing a condition of the lathe system based on an output from the current sensor, and wherein The diagnostic unit is configured to diagnose the condition of the lathe system using output information from the current sensor and information acquired by the control unit indicating a time elapsed since a switching point, which serves as a reference to any specified phase of the power conversion device. [2] Diagnostic device for the lathe system according to claim 1, wherein the output information of the current sensor is data obtained by performing a measurement of information for a waveform at a time of switching the power conversion device for each of several times after any period of time has elapsed since a switching time. [3] Diagnostic device for the lathe system according to claim 1 or 2, wherein the current sensor detects an overshoot waveform at the time of switching of the power conversion device and The diagnostic unit is configured to diagnose a condition of the lathe system using the captured overshoot waveform. [4] Diagnostic device for the lathe system according to claim 3, wherein the output information of the current sensor is data obtained by performing a measurement of overshoot waveform information at a time of switching of the power conversion device for each of several times after any period of time has elapsed since a switching time. [5] Diagnostic device for the lathe system according to one of claims 1 to 4, wherein signals from a temperature sensor and a humidity sensor are input into the diagnostic device. [6] Diagnostic device for the lathe system according to one of claims 1 to 5, wherein output information from the current sensor is acquired, a waveform of one or more periods is Fourier-transformed, a change in the amplitude of a predetermined frequency is extracted and a change amount or rate of change of the amplitude is calculated to perform a diagnosis of a condition. [7] Diagnostic device for the lathe system according to one of claims 1 to 6, wherein data for an Nth period are extracted from the output information of the current sensor and a diagnosis is performed on the data. [8] Diagnostic device for the lathe system according to any one of claims 1 to 7, wherein the diagnostic device is installed in an identical microcomputer as the control unit of the power conversion device. [9] Diagnostic device for the lathe system according to any one of claims 1 to 7, further comprising: an extraction unit (5) configured to extract data used for diagnostics from the output information of the current sensor; wherein the extraction unit is installed in an identical microcomputer as the control unit of the power conversion device and the diagnostic unit is installed outside the microcomputer. [10] Diagnostic device for the lathe system according to one of claims 1 to 7, wherein the diagnostic device is installed outside the power conversion device. [11] Lathe system comprising the following: a lathe (3), a power conversion device (7) and a cable that electrically connects the lathe and the power conversion device; wherein The power conversion device contains the following: a current sensor (10a, 10b) configured to measure a current supplied to the lathe; and a control unit configured to set a switching state based on an output from the current sensor; wherein The lathe system further includes the following: a diagnostic unit (6) configured to diagnose a condition of the lathe system based on an output from the current sensor, and wherein The diagnostic unit is configured to diagnose a condition using output information from the current sensor and information acquired by the control unit indicating a time elapsed since a switching point, which serves as a reference to an arbitrarily defined phase of the power conversion device. [12] Lathe system according to claim 11, wherein the output information of the current sensor is data obtained by asynchronously measuring information for a waveform at a time of switching of the power conversion device for each of several times after any period of time has elapsed since a switching time, with respect to switching an inverter. [13] Lathe system according to one of claims 11 or 12, wherein the current sensor detects an overshoot waveform at the time of switching of the power conversion device and The diagnostic unit is configured to diagnose a condition of the lathe system using the captured overshoot waveform. [14] Diagnostic method for a lathe system comprising a lathe (3), a power conversion device (7) and a cable electrically connecting the lathe and the power conversion device, wherein the diagnostic method comprises: Detect from an output of a current sensor (10a, 10b) provided and configured in the power conversion device to measure a current supplied to the lathe; Setting a switching state of the power conversion device by a control unit of the power conversion device based on the output of the current sensor; Diagnosing the condition of the lathe system by a diagnostic unit of a diagnostic device of the power conversion device based on the output of the current sensor; and Diagnosing the condition of the lathe system by the diagnostic unit using output information from the current sensor and information acquired by the control unit indicating a time elapsed since a switching point, which serves as a reference to an arbitrarily defined phase of the power conversion device. [15] Diagnostic method for the lathe system according to claim 14, wherein output information from the current sensor is acquired, a waveform of one or more periods is Fourier-transformed, a change in the amplitude of a predetermined frequency is extracted and a change amount or rate of change of the amplitude is calculated to perform a diagnosis of a condition. [16] Diagnostic method for the lathe system according to one of claims 14 or 15, wherein data for an Nth period are extracted from the output information of the current sensor and a diagnosis is performed on the data. [17] Diagnostic method for the lathe system according to any one of claims 14 to 16, further comprising: Extracting multiple data points from the current sensor's output information for different times after any given period has elapsed since a switching point; and Performing a diagnosis based on the data. [18] Diagnostic method for the lathe system according to any one of claims 14 to 17, further comprising: Extracting data from the output information of the current sensor for an overshoot waveform at the time of switching of the line converter device and Performing a diagnosis based on the data.

Images