WO2018158910A1

WO2018158910A1 - Diagnostic device and diagnostic method

Info

Publication number: WO2018158910A1
Application number: PCT/JP2017/008303
Authority: WO
Inventors: 哲司加藤; 牧　晃司; 永田　稔
Original assignee: 株式会社日立製作所
Priority date: 2017-03-02
Filing date: 2017-03-02
Publication date: 2018-09-07
Also published as: TW201835598A; TWI665458B

Abstract

Provided is a diagnostic device for highly accurately diagnosing a rotating machine even when a diagnosis is based on a small amount of current data. This diagnostic device is provided with a current measurement unit for measuring current flowing through at least two locations in a rotating machine and a diagnostic unit for diagnosing, on the basis of current data output by the current measurement unit, the state of the rotating machine and a peripheral device electrically or mechanically connected to the rotating machine. The diagnostic unit creates a Lissajous curve distribution map by layering multiple periods of the two types of current data obtained by the current measurement unit and diagnoses the state of the rotating machine system from the results of evaluating the distribution map.

Description

Diagnostic device and diagnostic method

The present invention relates to a diagnostic apparatus and a diagnostic method.

If a rotating machine such as a motor (electric motor) or generator built into a production facility suddenly breaks down, unscheduled repair or replacement of the rotating machine is required, which requires a reduction in the operating rate of the production facility or a review of the production plan. Become. Similarly, even if a power conversion device or cable connected to the rotating machine breaks down, unplanned repair work or replacement work is required, and it is necessary to reduce the operating rate of the production facility or review the production plan.

In order to prevent sudden failure of the rotating machine system (rotating machine and its associated devices (cables, power converters)), the rotating machine system is appropriately stopped and diagnosed offline, so that the deterioration condition can be grasped and It can be prevented to some extent. However, it is necessary to stop the rotating machine system for off-line diagnosis, leading to a reduction in the operating rate of the production facility. Some types of deterioration become apparent only when voltage is applied. Therefore, there is a need for diagnosing the state of the rotating machine based on the information on the current of the rotating machine system.

Non-patent document 1 is known as a conventional technique related to diagnosis based on current information of a rotating machine system. In Non-Patent Document 1, a method called Motor Current Signature Analysis (MCSA) is used to determine damage to rotor bars, rotor eccentricity, stator core damage, winding shorts, bearing deterioration, etc. The diagnosis can be made by detecting a specific frequency spectrum.

Also, in Patent Document 1, especially in bearing diagnosis, a method of acquiring vibration sensor data at two locations and judging an abnormality from a change in the trajectory inclination or radius of a Lissajous figure drawn with the instantaneous value of each sensor's data as an axis. Is disclosed.

JP 2000-258305 A

However, Non-Patent Document 1 and Patent Document 1 have the following problems. In the technique disclosed in Non-Patent Document 1, it is necessary to detect a specific frequency spectrum. To detect a specific frequency spectrum with high accuracy, it is expensive for diagnosis corresponding to long-time measurement at a high sampling rate. A new data logger is necessary, and the increase in diagnostic costs has been a problem. In addition, when the motor driving conditions change during long-time data measurement, the specific frequency spectrum may appear unintentionally, and there has been a problem of false alarms. Similarly, when the motor driving conditions change during long-term data measurement, the fundamental frequency changes and a spectrum appears at a frequency different from the expected frequency, causing a problem of misreporting.

Further, the technique disclosed in Patent Document 1 is a diagnosis based on vibration sensor information, and it is necessary to attach a vibration sensor to a position sensitive to a motor failure, and there is a problem that an installation place is limited. . In addition, a diagnostic sensor for diagnosis and an expensive data logger having a sampling rate sufficient to obtain a trajectory of a Lissajous figure are required, and an increase in diagnostic cost is a problem.

The object of the present invention is to provide a diagnostic apparatus that performs highly accurate diagnosis even when data measurement is performed in a short time, with a low sampling rate and with a general-purpose device.

In order to solve the above-described problems and achieve the object of the present invention, the following configuration is provided. That is, the diagnostic device of the present invention includes a current measurement unit that measures current flowing in at least two locations of the rotating machine, and a classification unit that classifies current data measured by the current measurement unit for each command value information of the power converter. A distribution of Lissajous figures obtained by superimposing a plurality of periods of current data of at least two phases classified by the classification unit (defined here as a distribution in which each point is not connected by a line in time series but plotted as a collection of points. Compared with the Lissajous figure's trajectory in that they are not connected in chronological order) and the Lissajous figure's distribution set as normal in advance, the status of the rotating machine or power converter is diagnosed from the change in the Lissajous figure's distribution A diagnostic unit is provided.

Further, the diagnostic method of the present invention measures the current flowing in at least two places of the rotating machine, classifies the measured current data for each command value information of the power converter, and classifies at least two phases and a plurality of cycles. Create a Lissajous figure distribution by superimposing current data, compare the created distribution with the Lissajous figure distribution set in advance as a normal state, and change the surroundings of the rotating machine or the power converter connected to the rotating machine Diagnose device status. Further details of the diagnostic method apparatus configuration and diagnostic method will be described in the detailed description.

According to the present invention, an expensive data logger is unnecessary, and the state of the rotating machine system can be diagnosed even when the driving condition of the motor changes.

Configuration diagram of diagnosis apparatus according to embodiment 1 The block diagram of the diagnostic apparatus of a conventional method. The figure which shows the example of a U-phase current waveform. The figure which shows the current waveform of the U phase of a normal state, and a W phase. The figure which shows the current waveform of the U phase of a deterioration state, and a W phase. The figure which shows distribution (high-speed sampling) of the Lissajous figure of a normal state and a deterioration state. The figure which shows the distribution (low-speed sampling) of the Lissajous figure of a normal state and a deterioration state. The figure which shows distribution of the Lissajous figure of the normal state and deterioration state classified by each control command. The figure which shows the diagnostic method by machine learning. The figure which shows the diagnostic result by Example 1. FIG. The block diagram by Example 2. FIG. FIG. 6 is a configuration diagram according to a third embodiment. The block diagram by Example 4. FIG. FIG. 6 is a configuration diagram according to a fifth embodiment.

Hereinafter, modes for carrying out the present invention (hereinafter referred to as “examples”) will be described with reference to the drawings as appropriate. Further, the following is merely an example of the embodiment, and the scope of the present invention is not intended to be limited to the following example.

The failure of a rotating machine system including a rotating machine such as an electric motor (motor) or a generator, a cable attached to the rotating machine, and a power conversion device, occurs in various parts, and a variety of failure factors. For example, insulation deterioration, bearing deterioration, short circuit, disconnection, water immersion, etc. can be considered. Moreover, the electric motor is often installed for a long time in a harsh environment, and a diagnostic technique according to the installation condition is required.

An example of a conventional diagnostic device 14 is shown in FIG. In a conventional diagnostic apparatus, one-phase current sensor information is acquired by the current measurement unit 11, and the diagnosis unit 12 performs diagnosis based on the value of the specific frequency spectrum of the spectrum obtained by Fourier transform. Since the change in the specific frequency spectrum is measured by Fourier transform, it is necessary to perform continuous measurement at a constant sampling rate. Therefore, it is necessary to increase the capacity of the memory for temporarily storing the measured data or increase the communication speed with the device for storing the data, and an expensive device is required. Moreover, since the change of the control signal is not assumed, there is a problem that the frequency of misreporting and misreporting is high.

The present inventors plot the intermittent sensor values for two phases among the load current values of the rotating machine on one plane, and visualize the state of the rotating machine as a data distribution map similar to one Lissajous figure. I examined that.

The Lissajous figure is a plane figure obtained by combining two waves. Since the current sensor data of the three-phase motor is shifted by 120 degrees from each other, when the two phases are combined, it becomes an inclined elliptical shape. Normally, the data is created by continuous data, but the inventors superimpose data for a plurality of frequencies obtained at predetermined times obtained intermittently, and use the resulting data distribution for evaluation. As a result, it is possible to diagnose the state of the rotating machine system with higher accuracy than intermittent data and to omit expensive equipment such as a data logger.

In order to solve the above problems, the diagnostic device of the present embodiment includes a current measuring unit that measures current flowing in at least two locations of the rotating machine, and a rotating machine and a rotating machine based on current data output from the current measuring unit. A diagnostic unit for diagnosing the state of a peripheral device electrically or mechanically connected. The diagnostic unit creates a Lissajous figure distribution map by overlapping a plurality of periods of two types of current data obtained by the current measurement unit, and diagnoses the state of the rotating machine system from the evaluation result of the distribution map.

Even in the case where the device as described above is not provided, the current flowing in at least two places of the rotating machine is measured, and the measured current data is classified for each command value information of the power conversion device. Create a Lissajous figure distribution map by overlaying the classified current data for multiple periods for each information, evaluate the created Lissajous figure distribution map, and diagnose the state of the rotating machine system based on the result Can do.

Also, the above diagnostic device can be incorporated into a rotating machine system. In particular, it is possible to reduce the number of parts by sharing a current sensor, a current measuring unit, and the like provided for rotating machine control in the power converter. Further, a plurality of rotating machines may be connected to the power conversion device.

In addition, by classifying and evaluating current data for each command value information of the power converter, it is possible to properly diagnose the state of the rotating machine system even when the driving conditions of the rotating machine change It is.

Specifically, it comprises one or a plurality of rotating machines, and a power converter that is electrically connected to the rotating machines and controls the current flowing through the rotating machines, and the power converters are provided at at least two locations of the rotating machines. A rotating machine system having a current measuring unit that measures a flowing current and a control unit that outputs a command value for controlling the rotating machine, and diagnoses the state of a device electrically or mechanically connected to the rotating machine And a classification unit that classifies the current data output from the current measurement unit and input to the diagnostic unit for each of the command values, and the diagnostic unit stores the current data of the rotating machine in at least two phases and Create a Lissajous figure distribution obtained by overlapping multiple periods.

Hereinafter, in the present embodiment, a current measurement unit that measures current flowing in at least two places of the rotating machine, and a classification unit that classifies the current data measured by the current measurement unit for each command value information of the power converter, Distribution of Lissajous figures obtained by superimposing a plurality of periods of current data of at least two phases classified by the classification unit (here, points are defined as distributions in which points are not connected by a line in time series but plotted as a collection of points. Compare the Lissajous figure's trajectory that is not connected in chronological order) and the Lissajous figure's distribution set as normal in advance, and diagnose the state of the rotating machine or power converter from the change in the Lissajous figure's distribution An example including a diagnosis unit will be described.

FIG. 1 shows a configuration diagram of the diagnostic apparatus of Example 1, and the diagnostic apparatus and the diagnostic method will be described. A common part with the description of FIG. 2 is omitted.

In Example 1, the power source 1, the cable 2, and the power conversion device 7 are electrically connected, and the power conversion device 7 outputs a three-phase AC voltage. The output of the three-phase AC voltage is controlled by adjusting the timing for operating the switching element of the inverter so that the rotation speed and torque of the motor have desired values. The control is determined based on control information set in advance and information on the current output from the inverter. The current information is acquired by the

current sensors

4a and 4b and the current measuring unit 9 and fed back to the control unit 8. The

Current information of a plurality of cycles of the

current sensors

4a and 4b of the current measuring unit is input to the classification unit, and the motor system is diagnosed by the distribution of the Lissajous figure plotted on the axis orthogonal to the

current sensors

4a and 4b.

The current information of the

current sensors

4a and 4b acquired by the current measuring unit does not necessarily have to be a constant sampling interval, and does not necessarily need to be a continuous measurement. The interval between the data acquisition of the current sensor 4a and the data acquisition of the current sensor 4b is preferably constant allowing a certain variation. By applying a measurement device that excels in real-time processing, a certain data acquisition interval that allows a certain variation can be obtained. For example, a measuring device using a microcomputer is an example.

Furthermore, by making the interval between data acquisition of the current sensor 4a and data acquisition of the current sensor 4b constant, there is no need for continuous measurement. For example, the current measurement unit 9 can be designed to store a certain amount of data in the memory of the microcomputer, insert a data communication process to the storage device, and then clear the memory to resume data storage. It is possible to make a system using a current sensor or a memory.

Hereinafter, as an example of a diagnostic method using the diagnostic apparatus of the first embodiment, when a control pattern does not change and the motor is operating at a constant fundamental frequency, a state where a spectrum is generated at a specific frequency is diagnosed. The method of doing this is described and the function of the diagnostic unit is explained.

Figure 3 shows the U-phase current waveform. FIG. 3a is a diagram showing a case where a frequency spectrum separated by 1 Hz from the fundamental frequency 50 Hz of the U-phase current appears. For example, in bearing deterioration, sideband waves of the fundamental frequency are generated due to the deterioration. In this case, as shown in FIG. 3b, the U-phase current appears as a waveform having a beat of 1 Hz period.

In order to accurately separate 50 Hz and 51 Hz components from this waveform with a conventional diagnostic device, it is necessary to continuously measure 20000 points of data for at least 100 seconds when measured at a frequency of 200 Hz. Therefore, there is a need for a measuring device equipped with a large-capacity memory that can measure more than 20000 points, or a measuring device that can write to a storage device at 200 Hz. In consideration of measurement errors, it is effective to increase either or both of the sampling speed and the measured data length. Therefore, it is necessary to apply a special and expensive measuring device in order to separate with high accuracy.

On the other hand, in this embodiment, diagnosis can be made by using any two types of current values, without increasing the sampling speed and the data length to be measured.

Fig. 4 shows normal U-phase and W-phase current waveforms, that is, a 50 Hz sine wave without undulations. The U-phase and W-phase currents are 120 degrees out of phase. On the other hand, FIG. 5 shows a conceptual diagram of U-phase and W-phase current waveforms when sidebands are generated, as in FIG. The sampling rate is 100 Hz and the data measurement time is 1 second.

As shown in FIGS. 4 and 5, based on the respective current waveforms obtained as a result of sampling at high speed, the U-phase current is plotted on the horizontal axis and the W-phase current is plotted on the vertical axis. The distribution is as follows. 6A is an example of a normal state based on FIG. 4, and FIG. 6B is a conceptual diagram when a sideband wave is generated due to deterioration of a rotating machine or the like based on FIG. 6A and 6B, it can be seen that in the deteriorated state, the distribution of the Lissajous figure is thicker than that in the normal state, and the thickness of the Lissajous figure distribution changes as the tendency of deterioration progresses.

FIG. 7 shows an example (FIG. 7a: normal, FIG. 7b: deterioration) in which Lissajous figure distributions are created for normal and deteriorated states using U-phase and W-phase currents obtained at a sampling rate of 4.975 Hz. The data length to be measured is 1000 points, and the data length is the same as in Figure 6. 6 and 7 show almost the same Lissajous figure distribution.

Therefore, it was possible to detect deterioration regardless of the sampling rate by evaluating the distribution of two-phase data. Regarding the difference between the deteriorated state and the normal state, the distribution of the Lissajous figure may be confirmed by human eyes, or the difference in the distribution of the Lissajous figure may be digitized and compared by a method such as machine learning. That is, according to the present embodiment, it is possible to easily detect the occurrence of undulation not only in the case of performing high-speed sampling but also in any case of a slow sampling frequency.

Note that when the sampling speed is slow, it is desirable to measure asynchronously with the fundamental frequency. Specifically, at a sampling rate having a period that is an integral multiple of the fundamental wave, a value of only a certain phase is always obtained, and there is a risk of false alarm in the case of deterioration in which a change appears only in a specific phase. Therefore, it is desirable that the sampling rate is a frequency different from an integral multiple of the fundamental wave period. As a result, a distribution map for diagnosis similar to the case where data is acquired at a high sampling rate can be acquired.

Also, in order to acquire high frequency information, it is desirable to measure asynchronously with the switching timing of the power converter.

Furthermore, if the Lissajous figure points are filled in time-sequential order with lines (Lissajous figure trajectory), the inside of the Lissajous figure distribution will be filled with lines, and the difference between the normal state and the deteriorated state may not be visible. It is desirable not to display the trajectory by connecting the lines in series order.

The diagnosis unit 10 outputs the above result. Means for transmitting the diagnosis result to the user can be selected as appropriate. Examples of the method for transmitting to the user include display of a display, lighting of a lamp, notification by e-mail, and the like. The contents are also (1) a method for displaying the distribution of Lissajous figures on the screen and allowing the user to determine whether or not it is compatible, (2) a method for quantifying the difference in the distribution of Lissajous figures in some way and telling the user (3) in advance A method of notifying the user when a predetermined threshold is exceeded can be considered.

As a method of quantifying the difference in the distribution of Lissajous figures in (2) above, machine learning can be applied. As an algorithm for machine learning, an algorithm that makes the difference between Lissajous figures clear should be selected. For example, a local subspace method can be cited. In the local subspace method, for all the points in the distribution of the Lissajous figure to be diagnosed, two closest points are selected from the distribution of the Lissajous figure defined as normal, and the straight line connecting the two points is diagnosed. In this method, the degree of deterioration is defined by the distance between target points.

In addition to the method of calculating the distance for all points to be diagnosed and digitizing the change in the distribution of the Lissajous figure, the average value of the distances of all the points to be diagnosed, the number of only the points of the specific phase, etc. An arbitrary evaluation method can be selected according to the variation error of the current waveform. When priority is given to the calculation speed, a clustering method such as vector quantization clustering or K-means clustering can be used. In addition, a technique called a deep neural network, which is a method for automatically finding feature quantities based on a large amount of data, can be applied.

Next, the case where the control pattern changes will be described. When the control pattern changes, the fundamental frequency, torque, etc. of the motor change. Therefore, the conventional method sometimes diagnoses this as an abnormality. Further, when learning is performed as a normal state including a state in which the control pattern is changed, a change due to deterioration may be overlooked and reported. Therefore, it is desirable to diagnose normality and abnormality for each control pattern.

In this example, the control command of the control unit and the current information were combined to improve the diagnostic accuracy. Since the diagnosis unit 10 diagnoses the state of the motor system based on the distribution of Lissajous figures drawn with the currents of the U phase and the W phase classified as the same state as the control command, the classification unit as shown in FIG. 6 was provided. Hereinafter, the function combining the classification unit 6 and the diagnosis unit 10 will be described.

As the control command, a voltage command value, a current command value, an excitation current command value, a torque current command value, a speed command value, a frequency command value, and the like can be arbitrarily selected from command values that can be output by the power converter 7 Can do. The classification unit 10 does not necessarily need to use all command values that can be output by the power conversion device 7, and can use only command values that are highly sensitive to the deterioration of the detection target. The command value with high sensitivity may be selected so that the difference between the deteriorated state and the normal state can be easily seen by comparing the distribution of the Lissajous figure in the deteriorated state with the distribution of the Lissajous figure in the normal state.

Hereinafter, the operation of the classification unit 6 and the diagnosis unit 10 based on the schematic diagram of the current waveform when the motor is driven under the two conditions of the control command A and the control command B having different voltage command values and frequency command values as control commands. Will be explained.

The classification unit 6 distributes the current information obtained from the current measurement unit 9 according to control command values (control command A and control command B) obtained from the control unit 8. Based on the distribution of the classification unit 6, the diagnosis unit 10 stores current information and command value information at intervals of about 1 minute, and creates a Lissajous figure distribution map for each control command value.

Figure 8 shows the distribution of normal and deteriorated Lissajous figures classified by each control command. 8a and 8c show the result of drawing the distribution of the Lissajous figure in the control command A, and FIGS. 8b and 8d in the control command B. FIG. As apparent from the comparison between FIG. 8a and FIG. 8b, in the case of the control command A and the control command B, the obtained Lissajous figure is different even in the normal state. Therefore, when the normal state of the control command A (FIG. 8a) is learned as normal, when the control command becomes a distribution of the Lissajous figure in the different control command B (FIG. 8b), the change is erroneously reported as a change due to deterioration. I will put it out. Therefore, misclassification can be suppressed by classifying the current waveform based on the control information in the classification unit.

In the classification unit, the current information acquired from the current measurement unit is divided into one associated with the control command A and one associated with the control command B. For each of the control command A and the control command B, a normal state and a deteriorated state are assigned. As a result of drawing the distribution of two types of Lissajous figures, comparing normal (FIG. 8a) and deterioration (FIG. 8c) in control command A, it can be seen that there is a difference in the distribution of Lissajous figures and that the deterioration can be detected. Further, when focusing on the control command B and comparing normal (FIG. 8b) and degradation (FIG. 8d) in the control command B, it can be seen that there is a difference in the distribution of the Lissajous figure and that the degradation can be detected.

Next, a method for quantifying the change of the Lissajous figure, taking as an example the case where the diagnosis object data X obtained at the time of the control command A is evaluated based on the result of learning that the control command A normal in FIG. Is described.

Fig. 9 shows learning data Y (control command A normal) in the vicinity of the data X to be diagnosed (control command A normal or degraded) in the Lissajous figure. Search for data X (control command A normal or degraded), diagnosis data Y1 (control command A normal) closest to the data to be diagnosed, and learning data Y2 (control command A normal) closest to the second The degree of abnormality was defined by the distance between the straight line connecting the data of Y1, Y2 and the data of the diagnosis target data X (control command A normal or deteriorated).

¡By performing the above contents for the measured data length, the degree of abnormality of the diagnosis target data group X1 to n (control command A normal or degraded) with respect to the learning data group Y can be quantified.

Fig. 10 shows the determination and evaluation results of the diagnosis results for the normal state and the abnormal state. An average value of the degree of abnormality of normal current data by the control command A and an average value of the degree of abnormality of current data when the device is deteriorated by the control command A are shown. The average value of the degree of abnormality increases due to deterioration. If a threshold value is set in advance by prior examination, an increase in the degree of abnormality, that is, the progress of deterioration can be displayed to the user.

In the first embodiment, the case where a spectrum is generated at a specific frequency has been described. However, even if the type of deterioration does not generate a spectrum at a specific frequency, some change in current occurs due to the load or impedance change of the motor due to the deterioration. Therefore, it can be detected by the diagnostic apparatus and diagnostic method of the first embodiment. Deterioration that does not appear as a change in the specific frequency spectrum is specifically deterioration other than the deterioration that appears as a peak at the specific frequency that can be detected by MCSA, that is, grease deterioration, thermal deterioration of insulation, and moisture absorption. Is assumed.

Further, according to the diagnostic apparatus of the first embodiment, it is possible to diagnose a motor system including a motor, a cable, a power converter, a load, and other devices electrically or mechanically connected to the motor. In motor systems that include peripheral devices connected to the motor, even if a peripheral device other than the motor fails or deteriorates, the current flowing through the motor changes due to changes in the impedance and load of those devices. Deterioration can be detected by a technique.

FIG. 11 is an example of a diagnostic system including a command unit 13. The command unit 13 sets a control command in the control unit 8 and changes it. In the first embodiment, the information of the control unit 8 is input to the classification unit and diagnosed. However, the information of the command unit 13 is input to the classification unit together with the control command value obtained from the control unit 8 or instead of the control command value. And may be diagnosed. As shown in FIG. 11, when information such as the presence of multiple types of loads to be rotated, changes in control command values according to time, etc. are stored in the command unit 13, the information of the command unit 13 is input to the classification unit, You may make a diagnosis. Thereby, diagnostic accuracy can be improved.

In Example 1 and Example 2, an example in which one motor 3 is connected to one power conversion device 7 has been described. In Example 3, as shown in FIG. 12, an example in which a plurality of motors 3 are connected to one power conversion device 7 will be described.

In FIG. 12, diagnosis is performed using the result of measuring the entire load current by the current sensor in the power converter. The diagnosis method is the same as when one rotating machine is connected, and the diagnosis data is diagnosed as a Lissajous figure distribution along with classification using control command values and reflection of command value information.

In addition, since the current signals of multiple rotating machines are diagnosed as one load current, the change in the current waveform of the deteriorated motor is diminished by the current waveforms of the other multiple motors. It is desirable to evaluate the change of Lissajous figure distribution by the spatial method.

In the first to third embodiments, the current sensor and the current measurement unit in the power converter are used. However, as shown in FIG. 13, the

current sensors

4a and 4b and the current measurement unit 9 prepared separately from the power converter are used. Even so, the effect of the present invention can be obtained. Since it is not necessary to support long-term data measurement at a high sampling rate, an inexpensive general-purpose measuring device can be used.

In the third embodiment, in FIG. 12, when a plurality of motors and one power conversion device are connected, the current sensor and the current measurement unit in the power conversion device are used. A current sensor may be provided. Moreover, it is good also as providing the current sensor independent of a power converter device in the position where a motor and a power converter device are connected, and measuring the whole load current.

In addition, when measuring the load currents of multiple motors with multiple current sensors, a classification unit / diagnostic unit can be provided for each rotating machine, or multiple sensor information can be diagnosed with a single classification unit / diagnostic unit. It is also possible to process.

Example 5 describes an example in which three current sensors are installed for each of a zero-phase current and a two-phase load current. In the first to fourth embodiments, the state of the motor system is diagnosed based on the information of the two current sensors, but as shown in FIG. 14, the distribution of the Lissajous figure is based on the information of three or more current sensors. You may get. In this case, when comparing the distributions, it is possible to obtain and evaluate a Lissajous figure distribution by a plurality of combinations of two selected from three or more current sensors.

Also, it is possible to evaluate changes in the distribution of Lissajous figures in three or more dimensional spaces using machine learning, for example, the local subspace method, by taking three or more current sensors as orthogonal axes.

The current value measured by the current sensor, in the case of a three-phase AC rotating machine, in addition to the U-phase, V-phase, and W-phase load currents, as well as the zero-phase current measured by clamping three phases Two or more currents can be set, such as two-phase current measured by clamping the selected two phases, leakage current measured by clamping both the winding start and end of the motor winding, and current flowing from the motor to the ground. By doing in this way, the position of a highly sensitive current sensor can be selected for each degradation mechanism. For example, bearing deterioration can be detected with high sensitivity at load current, and insulation deterioration can be detected with zero-phase current. What is necessary is just to select arbitrarily the electric current which raises diagnostic sensitivity with respect to the cause of deterioration, and to install a current sensor in the position.

DESCRIPTION OF SYMBOLS 1 Power supply 2 Cable 3

Motor

4a, 4b, 4c Current sensor 5 Extraction part 6 Classification | category part 7 Power converter 8 Control part 9 Current measurement part 10 Diagnosis part 11 Current measurement part 12 Diagnosis part 13 Command part 14 Conventional diagnostic apparatus

Claims

A current measuring unit that measures current flowing in at least two locations of the rotating machine;
A diagnostic unit for diagnosing the state of the rotating machine and peripheral devices electrically or mechanically connected to the rotating machine based on current data output from the current measuring unit;
A diagnostic device for a rotating machine system comprising:
The diagnostic unit creates a Lissajous figure distribution diagram by overlapping a plurality of periods of the two types of current data obtained by the current measurement unit, and diagnoses the state of the rotating machine system from the evaluation result of the distribution diagram. Diagnostic equipment for rotating machine system.
In the diagnostic apparatus for a rotating machine system according to claim 1,
A classification unit that classifies the current data for each command value information of a power conversion device that is electrically connected to the rotating machine and controls a current flowing through the rotating machine,
The diagnostic device for a rotating machine system, wherein the diagnostic unit uses at least two-phase current data classified by the classification unit.
In the diagnostic apparatus for a rotating machine system according to claim 1 or 2,
The diagnostic unit compares the created Lissajous figure distribution with a Lissajous figure distribution set as a normal state in advance, and diagnoses the presence or absence of an abnormality.
In the diagnostic apparatus for a rotating machine system according to claim 3,
The diagnostic device for a rotating machine system, wherein the diagnosis unit quantifies and diagnoses a change in the distribution of the created Lissajous figure with respect to the distribution of the Lissajous figure set as a normal state in advance.
In the diagnostic apparatus for a rotating machine system according to any one of claims 1 to 4,
The diagnostic device for a rotating machine system, wherein the current measuring unit is provided in a power converter that is electrically connected to the rotating machine and controls a current flowing through the rotating machine.
In the diagnostic apparatus for a rotating machine system according to any one of claims 1 to 5,
The diagnostic apparatus for a rotating machine system, wherein the current measuring unit measures current data at a frequency different from an integer multiple of a fundamental wave period of the rotating machine.
A diagnostic method for a rotating machine system comprising: one or a plurality of rotating machines; and a power conversion device that is electrically connected to the rotating machines and controls a current flowing through the rotating machines,
Measure the current flowing in at least two places of the rotating machine,
Classifying the measured current data for each command value information of the power converter,
For each of the command value information, create a Lissajous figure distribution map by overlapping the classified current data for a plurality of periods,
Diagnosing the state of the rotating machine system based on the created Lissajous figure distribution map;
A diagnostic method for a rotating machine system.
In the diagnostic method of the rotating machine system in Claim 7,
A diagnostic method for a rotating machine system, characterized by using at least two-phase current data.
In the diagnostic method of the rotating machine system in Claim 7,
The diagnosis includes comparing the created Lissajous figure distribution with a Lissajous figure distribution set in a normal state in advance.
In the diagnosis method of the rotating machine system according to claim 9,
The diagnostic method for a rotating machine system is characterized in that a change in the distribution of the created Lissajous figure with respect to the distribution of the Lissajous figure previously set as a normal state is quantified.
In the diagnostic method of the rotating machine system in Claim 7,
The method for diagnosing a rotating machine system, wherein the current data is measured at a frequency different from an integer multiple of a period of a fundamental wave of the rotating machine.
One or a plurality of rotating machines, and a power conversion device that is electrically connected to the rotating machines and controls a current flowing through the rotating machines,
The power converter is a rotating machine system having a current measuring unit that measures current flowing in at least two locations of the rotating machine, and a control unit that outputs a command value for controlling the rotating machine,
A diagnostic unit for diagnosing the state of equipment electrically or mechanically connected to the rotating machine;
A classification unit that classifies current data output from the current measurement unit and input to the diagnosis unit for each command value;
The diagnostic unit creates a Lissajous figure distribution obtained by superimposing current data of the rotating machine for at least two phases and a plurality of cycles.