Estimation device, estimation program, estimation method, motor bearing wear monitoring device, and canned motor pump

Abstract

This reduces the man-hours required to adjust the correspondence between the detection value of the detection coil and the amount of bearing wear. [Solution] The estimation devices 53,6 according to the present invention estimate the adjustment value of a motor bearing wear monitoring device 5 that monitors the wear state of bearings 32,33 supporting a rotating shaft 31, based on the detection signals of each of a plurality of detection coils C1 to C8. The estimation device includes storage units 54,64 that store trained thrust learning models M11,M21 that are machine-learned to output a thrust adjustment value when thrust input data including the rated current value of the motor 3, the length of the movable range, the number of turns of the detection coil, and the position of the rotor 36 relative to the stator 37 in the thrust direction is input; acquisition units 530,630 that acquire thrust input data; and estimation units 531,631 that input the thrust input data acquired by the acquisition units into the thrust learning model to estimate the thrust adjustment value.

Description

[Technical Field]

[0001] The present invention relates to an estimation device, an estimation program, an estimation method, a motor bearing wear monitoring device, and a canned motor pump. [Background technology]

[0002] Canned motor pumps have a structure in which the pump and motor are integrated, preventing leakage of the handled fluid. Generally, the rotating parts of a canned motor pump (rotor, rotating shaft, bearings, and impeller) are sealed in a can filled with the handled fluid. Therefore, the internal structure of a canned motor pump cannot be monitored visually from the outside. Accordingly, in order to efficiently operate a canned motor pump with such a structure, a device for monitoring the wear condition of the bearings (hereinafter referred to as "monitoring device") is used (see, for example, Patent Document 1).

[0003] The monitoring device (motor bearing wear monitoring device) disclosed in Patent Document 1 monitors the radial and thrust displacement of the rotor (rotating shaft) caused by bearing wear by measuring the change in magnetic flux during rotor rotation using detection coils attached to both ends in the longitudinal direction of the stator. In this method, it is necessary to adjust the correspondence between the amount of bearing wear (mechanical displacement of the rotor) and the output value of the detection coil so that the output value (voltage value) of the detection coil indicates the amount of bearing wear.

[0004] Normally, bearings have some play. Therefore, for example, in the thrust direction, the rotor can displace by a length corresponding to the bearing play (hereinafter referred to as "play length") without the bearing wearing down. When the rotor displaces beyond the play length, the bearing is worn. Consequently, the output value of the detection coil when the rotor is displaced by the play length becomes a threshold indicating that the bearing is not worn. The output value of the detection coil is approximately proportional to the amount of rotor displacement. Therefore, the output value corresponding to the limit of bearing wear can be calculated from the output value corresponding to the play length. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 10-080103 [Overview of the project] [Problems that the invention aims to solve]

[0006] Here, the monitoring device detects the voltage induced in the detection coil by the rotation of the motor. Therefore, the adjustment is performed with the canned motor pump running under predetermined operating conditions to pump the fluid being handled (e.g., water). During this process, the operator moves the rotor as far as possible to the front and rear within the play length range in the thrust direction, obtaining the threshold values ​​(output values) at the front and rear, as well as the output value at an intermediate position between the front and rear. In other words, the adjustment requires multiple rotor movements and fluid pumping operations by the canned motor pump. Thus, the adjustment requires manual mechanical operation, and the process is time-consuming.

[0007] The present invention aims to provide an estimation device, estimation program, estimation method, motor bearing wear monitoring device, and canned motor pump that can reduce the man-hours required to adjust the correspondence between the output value of a detection coil and the amount of bearing wear. [Means for solving the problem]

[0008] An estimation device in one embodiment of the present invention is an estimation device that estimates an adjustment value for a motor bearing wear monitoring device that monitors the wear state of a bearing supporting the rotation axis of a canned motor pump, based on the detection signals of each of a plurality of detection coils that detect a change in magnetic flux corresponding to a change in the position of the rotor relative to the stator of the motor, wherein each of the plurality of detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, the plurality of detection coils include a plurality of thrust detection coils that detect the change in magnetic flux in the thrust direction of the rotation axis, the plurality of thrust detection coils include a pair of first thrust detection coils and another pair of second thrust detection coils, in the thrust direction, the first thrust detection coil is located at one end of the stator, and the second thrust detection coil is located at the other end of the stator, the adjustment value includes a thrust adjustment value in the thrust direction, and the thrust adjustment value is the first thrust detection coil when the rotor is located at a first position, a second position, and a third position relative to the stator, respectively. The system includes a thrust determination criterion for the wear state in the thrust direction, which can be calculated based on the voltage value of a first thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, the voltage value of a second thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, wherein the first position is the position in which the rotor has moved the maximum distance to one direction relative to the stator within the mechanically movable range of the rotating shaft relative to the bearing in the thrust direction, the second position is the position in which the rotor has moved the maximum distance to the other direction relative to the stator within the movable range, and the third position is an intermediate position between the first and second positions, and a storage unit that stores a trained thrust learning model which has been machine-learned to output the thrust adjustment value when thrust input data is input including the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction.The system comprises an acquisition unit that acquires the thrust input data, and an estimation unit that inputs the thrust input data acquired by the acquisition unit into the thrust learning model to estimate the thrust adjustment value.

[0009] The estimation program in one embodiment of the present invention causes a computer to function as the estimation device described in the above embodiment.

[0010] An estimation method in one embodiment of the present invention is an estimation method for estimating an adjustment value of a motor bearing wear monitoring device that monitors the wear state of a bearing supporting the rotation axis of a canned motor pump, based on the detection signals of each of a plurality of detection coils that detect a change in magnetic flux corresponding to a change in the position of the rotor relative to the stator of the motor, wherein each of the plurality of detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, the plurality of detection coils include a plurality of thrust detection coils that detect the change in magnetic flux in the thrust direction of the rotation axis, the plurality of thrust detection coils include a pair of first thrust detection coils and another pair of second thrust detection coils, in the thrust direction, the first thrust detection coil is located at one end of the stator, and the second thrust detection coil is located at the other end of the stator, and the estimation device determines the rated current value of the motor, the length of the mechanically movable range of the rotation axis relative to the bearing in the thrust direction, and the number of turns of the detection coil The estimation device includes a storage unit that stores a trained thrust learning model that has been machine-trained to output a thrust adjustment value when thrust input data is input, including the position of the rotor relative to the stator in the thrust direction, and the estimation device includes an acquisition step of acquiring the thrust input data, and an estimation step of inputting the thrust input data acquired in the acquisition step into the thrust learning model to estimate the thrust adjustment value, wherein the thrust adjustment value includes a thrust determination criterion for the wear state in the thrust direction, which is generated based on the voltage value of a first thrust composite signal obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is located at a first position, a second position, and a third position relative to the stator, respectively, and the voltage value of a second thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, and the first position is within the movable range,The first position is the position where the rotor is moved to its maximum extent in one direction relative to the stator; the second position is the position where the rotor is moved to its maximum extent in the other direction relative to the stator within the movable range; and the third position is an intermediate position between the first and second positions.

[0011] A motor bearing wear monitoring device in one embodiment of the present invention is a motor bearing wear monitoring device that monitors the wear state of a bearing supporting the rotation axis of a rotor of a canned motor pump based on the detection signals of each of a plurality of detection coils that detect a change in magnetic flux corresponding to a change in the position of the rotor relative to the stator of the motor, wherein each of the plurality of detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, the plurality of detection coils includes a plurality of thrust detection coils that detect the change in magnetic flux in the thrust direction of the rotation axis, the plurality of thrust detection coils includes a pair of first thrust detection coils and another pair of second thrust detection coils, in the thrust direction, the first thrust detection coil is located at one end of the stator, and the second thrust detection coil is located at the other end of the stator, the adjustment value includes a thrust adjustment value in the thrust direction, and the thrust adjustment value is the output from each of the first thrust detection coils when the rotor is located at a first position, a second position and a third position relative to the stator, respectively. The system includes a thrust determination criterion for the wear state in the thrust direction, which is generated based on the voltage value of a first thrust composite signal obtained by combining the detected signals, the voltage value of a second thrust composite signal obtained by combining the detected signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, wherein the first position is the position in which the rotor has moved the maximum distance to the 1stator relative to the stator within the mechanically movable range of the rotating shaft relative to the bearing in the thrust direction. The second position is the position in which the rotor moves to the maximum extent possible in the other direction relative to the stator within the movable range, and the third position is an intermediate position between the first position and the second position. The system includes a storage unit that stores a trained thrust learning model that has been machine-learned to output the thrust adjustment value when thrust input data is input, which includes the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction, and an acquisition unit that acquires the thrust input data.An estimation unit that inputs the thrust input data acquired by the acquisition unit into the thrust learning model and estimates the thrust adjustment value.

[0012] A canned motor pump according to an embodiment of the present invention includes a rotor, a stator that rotates the rotor, a rotating shaft to which the rotation of the rotor is transmitted, a bearing that rotatably supports the rotating shaft, a plurality of detection coils that are attached to the stator and detect a change in magnetic flux corresponding to a change in the position of the rotor with respect to the stator, and a motor bearing wear monitoring device that monitors the wear state of the bearing based on an output signal of the detection coil. The motor bearing wear monitoring device is the motor bearing wear monitoring device described in the above embodiment.

Advantages of the Invention

[0013] The present invention can reduce the man-hour for adjusting the correspondence relationship between the output value of the detection coil and the wear amount of the bearing.

Brief Description of the Drawings

[0014] [Figure 1] It is a side view of the canned motor pump showing an embodiment of the canned motor pump according to the present invention. [Figure 2] It is a schematic cross-sectional view of the motor part showing a longitudinal cross-section of the motor part of the canned motor pump of FIG. 1. [Figure 3] It is a partially enlarged schematic cross-sectional view showing part A of the motor part of FIG. 2. [Figure 4] It is a partially enlarged schematic cross-sectional view showing part B of the motor part of FIG. 2. [Figure 5] It is a partially enlarged perspective view showing part C of the motor part of FIG. 2. [Figure 6] It is a functional block diagram showing an embodiment of the motor bearing wear monitoring device according to the present invention. [Figure 7] It is a schematic perspective view of the stator core of the canned motor pump of FIG. 1 showing the arrangement of the detection coils of the motor bearing wear monitoring device of FIG. 6. [Figure 8]Figure 7 is a schematic diagram showing an example of the detection signal from the detection coil. [Figure 9] Figure 6 is a functional block diagram of the offset processing unit of the motor bearing wear monitoring device. [Figure 10] Figure 6 is a schematic diagram showing an example of the information stored in the memory unit of the motor bearing wear monitoring device. [Figure 11] Figure 6 is a flowchart illustrating an example of the operation of a motor bearing wear monitoring device. [Figure 12] This flowchart shows an example of the thrust adjustment value estimation process included in the operation shown in Figure 11. [Figure 13] This flowchart shows an example of the radial adjustment value estimation process included in the operation shown in Figure 11. [Figure 14] Figure 11 shows a flowchart illustrating an example of the judgment criterion setting process included in the operation. [Figure 15] This is a functional block diagram of the estimation device, showing another embodiment of the estimation device according to the present invention. [Figure 16] Figure 15 is a schematic diagram showing an example of the information stored in the memory unit of the estimation device. [Figure 17] Figure 15 is a flowchart illustrating an example of the operation of the estimation device. [Figure 18] Figure 17 shows a flowchart illustrating an example of the thrust adjustment value estimation process included in the operation. [Figure 19] Figure 17 shows a flowchart illustrating an example of the radial adjustment value estimation process included in the operation. [Modes for carrying out the invention]

[0015] This invention estimates the adjustment values ​​used for adjusting the motor bearing wear monitoring device of a canned motor pump using a pre-trained model. The machine learning of this model utilizes data (big data) accumulated from past adjustments. As a result, this invention allows for the estimation of adjustment values ​​and adjustment of the motor bearing wear monitoring device without human intervention, for example, during initial adjustments before shipment of a canned motor pump or when driving conditions are changed. Details of each term will be described later.

[0016] Embodiments of the canned motor pump (hereinafter referred to as "the pump"), the motor bearing wear monitoring device (hereinafter referred to as "the device"), the estimation device (hereinafter referred to as "the estimation device"), the estimation method (hereinafter referred to as "the estimation method"), and the estimation program (hereinafter referred to as "the estimation program") according to the present invention are described below. In the following description, the drawings will be referenced as appropriate. In each drawing, the same reference numerals are used for the same members and elements, and redundant explanations are omitted. In addition, the dimensional ratios of each element may be exaggerated for the sake of explanation, and are not limited to the ratios shown in each drawing.

[0017] ●Canned Motor Pump● First, the first embodiment of this estimation device will be described below, using the case where the device of the pump is equipped with this estimation device as an example.

[0018] ● Configuration of the canned motor pump Figure 1 is a side view of the pump, showing an embodiment of this pump. For the sake of explanation, the upper half of pump 1 is shown as a cross-sectional view in the diagram.

[0019] This pump 1 is a pump with a structure that prevents leakage of the liquid being handled. This pump 1 comprises a pump unit 2, a motor unit 3, an adapter 4, and this device 5.

[0020] The configurations of the pump unit 2, motor unit 3, and adapter 4 of this pump 1 are common to those of a known canned motor pump. Therefore, in the following description, only a general overview of the configurations of the pump unit 2, motor unit 3, and adapter 4 will be provided, and detailed explanations will be omitted.

[0021] In the following description, "front direction" refers to the direction in which the pump unit 2 is located relative to the motor unit 3, and "rear direction" refers to the direction in which the motor unit 3 is located relative to the pump unit 2. The front direction is an example of one direction in the present invention. The rear direction is an example of another direction in the present invention.

[0022] Pump unit 2 sucks in and discharges the liquid being handled. Pump unit 2 comprises a housing 20, an impeller 21, a pump chamber 22, a suction pipe section 23, and a discharge pipe section 24. The housing 20 forms the pump chamber 22 which houses the impeller 21, the suction pipe section 23 which is the path for the liquid being handled that is sucked into the pump chamber 22, and the discharge pipe section 24 which is the path for the liquid being handled that is discharged from the pump chamber 22. The pump chamber 22 is in communication with the suction pipe section 23 and the discharge pipe section 24.

[0023] The motor unit 3 is driven under rated conditions (for example, rated frequency: 60Hz, rated voltage: 200V, rated current: 10A) to rotate the impeller 21 of the pump unit 2. The motor unit 3 comprises a housing 30, a rotating shaft 31, two bearings 32, 33, two thrust washers 34, 35, a rotor 36, a stator 37, a can 38, and terminal terminals 39. The motor unit 3 is an example of a motor in the present invention.

[0024] Figure 2 is a schematic cross-sectional view of the motor unit 3, showing a longitudinal section of the motor unit 3.

[0025] The housing 30 contains the stator 37 and the can 38 in a liquid-tight manner.

[0026] The rotating shaft 31 rotates together with the rotor 36 and transmits rotational power to the impeller 21. The rotating shaft 31 is cylindrical in shape. The rotating shaft 31 is inserted into the rotor 36 and fixed to the rotor 36. The front end of the rotating shaft 31 protrudes into the pump chamber 22 (see Figure 1), and the impeller 21 is attached to this front end. The rotating shaft 31 is equipped with cylindrical sleeves 31a and 31b that protect the front and rear portions of the rotating shaft 31.

[0027] In the following explanation, "thrust direction" refers to the thrust direction of the rotation axis 31, "radial direction" refers to the radial direction of the rotation axis 31, and "circumferential direction" refers to the circumferential direction of the rotation axis 31.

[0028] Bearing 32 is positioned in the front direction of the rotor 36 and rotatably supports the rotating shaft 31. Bearing 33 is positioned in the rear direction of the rotor 36 and rotatably supports the rotating shaft 31. Bearings 32 and 33 are, for example, sliding bearings. A thrust washer 34 is installed on the rotating shaft 31 between bearing 32 and the rotor 36 and restricts the movement of the rotating shaft 31 in the front direction. A thrust washer 35 is installed on the rotating shaft 31 between bearing 33 and the rotor 36 and restricts the movement of the rotating shaft 31 in the rear direction.

[0029] Figure 3 is a partially enlarged schematic cross-sectional view showing part A of the motor unit 3 in Figure 2. Figure 4 is a partially enlarged schematic cross-sectional view showing part B of the motor section 3 in Figure 2.

[0030] In the thrust direction, a gap Ss1 of length "Ls1" is formed between the bearing 32 and the thrust washer 34. In the thrust direction, a gap Ss2 of length "Ls2" is formed between the bearing 33 and the thrust washer 35. The lengths "Ls1" and "Ls2" of these gaps Ss1 and Ss2 define the mechanically movable range of the rotating shaft 31 relative to the bearings 32 and 33 in the thrust direction, that is, the play of the bearings 32 and 33 in the thrust direction. In other words, the length of this movable range (hereinafter referred to as "thrust play length") is expressed as the sum of the lengths "Ls1" and "Ls2", "Ls1 + Ls2".

[0031] When bearing 32 ideally supports the rotating shaft 31, a gap Sr1 of length "Lr1" is formed between the rotating shaft 31 and bearing 32 in the radial direction. When bearing 33 ideally supports the rotating shaft 31, a gap Sr2 of length "Lr2" is formed between the rotating shaft 31 and bearing 33 in the radial direction. Length "Lr1" is approximately the same as length "Lr2". These gaps Sr1 and Sr2 define the mechanically movable range of the rotating shaft 31 relative to bearings 32 and 33 in the radial direction, that is, the play of bearings 32 and 33 in the radial direction. That is, when bearings 32 and 33 ideally support the rotating shaft 31, the length of this movable range (hereinafter referred to as "radial play length") is expressed by length "Lr1" or length "Lr2". Here, when bearings 32 and 33 ideally support the rotating shaft 31, the rotor 36 is located at the center position in the radial direction. The "center position" is the position where, in a thrust direction view, the center of the rotation axis 31 coincides with the centers of the bearings 32 and 33.

[0032] In the following explanation, the drawings primarily referenced will return to Figures 1 through 4. The rotor 36 rotates due to the rotating magnetic field generated in the stator 37. The rotor 36 has a cylindrical shape. The rotor 36 is equipped with a plurality (28 in this embodiment) of rod-shaped rotor bars 36a embedded at equal intervals in the outer peripheral edge of the rotor 36 in the circumferential direction. When the bearings 32 and 33 are not worn, the rotor 36 is positioned in its initial position relative to the stator 37. In this embodiment, the "initial position" is the position where the center of the stator 37 and the center of the rotor 36 coincide in the thrust direction and the radial direction.

[0033] In the following description, "front position" means the position where the rotor 36 has moved the maximum distance in the forward direction within the range of motion in the thrust direction. That is, the front position is the position where the rotor 36 has moved a length "Ls1" forward from its initial position. Also, "rear position" means the position where the rotor 36 has moved the maximum distance in the rear direction from its initial position within the range of motion in the thrust direction. That is, the rear position is the position where the rotor 36 has moved a length "Ls2" rear from its initial position. "Radial position" is the position where the rotor 36 has moved the maximum distance in the radial direction from its center position within the range of motion in the radial direction. That is, the radial position is the position where the rotor 36 has moved a length "Lr1" or "Lr2" radially from its center position (initial position). The front position is an example of the first position in this invention. The rear position is an example of the second position in this invention. An intermediate position between the front position and the rear position is an example of the third position in this invention. The radial position is an example of the fourth position in this invention.

[0034] The stator 37 generates a rotating magnetic field that rotates the rotor 36. The stator 37 has a roughly cylindrical shape. The stator 37 comprises a stator core 37a and a plurality of motor windings 37b.

[0035] Figure 5 is a partially enlarged perspective view showing section C of the motor unit 3 in Figure 2. In the following explanation, Figures 1 to 4 will be referred to as appropriate, along with Figure 5.

[0036] The stator core 37a holds the motor windings 37b. The stator core 37a has a cylindrical shape. The stator core 37a is provided with a plurality of teeth 37c, slots 37d, and a plurality of notches 37e.

[0037] The teeth 37c form slots 37d through which the motor windings 37b are inserted. In the circumferential direction, the teeth 37c are arranged at equal intervals on the inner circumferential surface of the stator core 37a. The motor windings 37b are inserted into the slots 37d and connected to a power supply device (not shown), such as an inverter, via terminal terminals 39. The notches 37e are arranged at equal intervals (90° intervals) in the circumferential direction at both ends of the teeth 37c in the thrust direction.

[0038] The can 38 liquid-tightly houses the rotating shaft 31, bearings 32 and 33, thrust washers 34 and 35, and rotor 36. The can 38 is cylindrical in shape. A portion of the handling fluid introduced from the suction pipe section 23 is introduced into the can 38 and used for cooling and lubricating the bearings 32 and 33 and the motor section 3, before being discharged into the discharge pipe section 24.

[0039] The adapter 4 is connected to the rear end of the pump unit 2 and the front end of the motor unit 3, thus connecting the pump unit 2 and the motor unit 3.

[0040] The device 5 monitors the amount of wear (wear condition) of bearings 32 and 33 by detecting changes in magnetic flux corresponding to changes in the mechanical position of the rotor 36 relative to the stator 37. The specific configuration of the device 5 will be described later.

[0041] ●Motor bearing wear state estimation device● ●Configuration of the motor bearing wear state estimation device Next, the configuration of the device 5 will be described. Figures 1 to 4 will be referenced as appropriate in the following description of the device 5.

[0042] Figure 6 is a functional block diagram showing an embodiment of the device 5.

[0043] The device 5 comprises eight detection coils C1, C2, C3, C4, C5, C6, C7, C8, a connection section 50, signal processing circuits 51a, 51b, 51c, 51d, an A / D converter 52, a control unit 53, a storage unit 54, a display unit 55, and an offset processing unit 56. The A / D converter 52 and the control unit 53 are configured, for example, by a microcomputer. The control unit 53 and the storage unit 54 are examples of this estimation device. In other words, the pump 1 and the device 5 are equipped with this estimation device.

[0044] In addition, in the present invention, the motor unit 3 may include detection coils C1 to C8.

[0045] Figure 7 is a schematic perspective view of the stator core 37a showing the arrangement of detection coils C1 to C8.

[0046] Each of the detection coils C1 to C8 detects a change in magnetic flux corresponding to the change in position (displacement) of the rotor 36 relative to the stator 37, generates a detection signal indicating the change in magnetic flux, and outputs the same detection signal. The rotor 36 is displaced in the radial and thrust directions together with the rotating shaft 31 according to the amount of wear in the radial and thrust directions of the bearings 32 and 33. That is, the amount of displacement of the rotor 36 can be considered as the amount of wear of the bearings 32 and 33. Therefore, this device 5 can detect the amount of wear of the bearings 32 and 33 by acquiring a detection signal indicating the amount of displacement of the rotor 36 using the detection coils C1 to C8. Each of the detection coils C1 to C8 is shaped like a flat bobbin. Each of the detection coils C1 to C8 is fitted into the corresponding notch 37e.

[0047] In the circumferential direction, detection coils C1 to C4 are mounted at equal intervals (90° intervals) on the front end of the tooth portion 37c. In the circumferential direction, detection coil C1 is positioned to face detection coil C3 at a position of "180°", and detection coil C2 is positioned to face detection coil C4 at a position of "180°". On the other hand, in the circumferential direction, detection coils C5 to C8 are mounted at equal intervals (90° intervals) on the rear end of the tooth portion 37c. In the circumferential direction, detection coil C5 is positioned to face detection coil C7 at a position of "180°", and detection coil C6 is positioned to face detection coil C8 at a position of "180°". That is, in the radial direction, detection coils C1, C2, C5, and C6 are positioned to face their corresponding detection coils C3, C4, C7, and C8.

[0048] In this invention, detection coils C1, C2, C5, and C6 only need to be arranged so as to face their corresponding detection coils C3, C4, C7, and C8, and their positional relationship is not limited to an exact "180°" position. That is, for example, due to manufacturing tolerances or the shape of the stator 37, their positional relationship in the circumferential direction may be off by a few degrees (e.g., 1° to 5°) from "180°".

[0049] Furthermore, in the present invention, the angle between detection coil C1 (detection coil C3) and detection coil C2 (detection coil C4) in the circumferential direction can be appropriately set between "0° and 180°" and is not limited to "90°". The same arrangement applies to detection coil C5 (detection coil C7) and detection coil C6 (detection coil C8).

[0050] Figure 8 is a schematic diagram showing an example of a detected signal.

[0051] In the figure, the horizontal axis represents the rotation angle of the rotor 36, and the vertical axis represents the output voltage (signal level) of the detection coils C1 to C8. The detection signals from the detection coils C1 to C8 include a waveform corresponding to the change in the main magnetic flux of the motor unit 3 (hereinafter referred to as the "fundamental wave component") and a waveform corresponding to the change in magnetic flux generated by the induced current flowing through the rotor bars 36a of the rotor 36 (hereinafter referred to as the "harmonic component"). The fundamental wave component is generated by the drive voltage of the motor unit 3, and its frequency is the same as the drive frequency. The harmonic component is generated by the induced current flowing through the rotor bars 36a, and its frequency is determined by the rotation of the rotor 36 and the number of rotor bars 36a. That is, for example, under the following conditions (drive frequency: 60Hz, number of rotor bars 36a: 28), each of the detection coils C1 to C8 detects the change in magnetic flux due to the rotor bars 36a 28 times during one rotation of the rotor 36. Therefore, the frequency of the harmonic component is 60Hz × 28 = 1.68kHz. Thus, the fundamental wave component is determined based on the drive frequency, and the harmonic components are determined based on the rotation of the rotor 36, the drive frequency, and the number of rotor bars 36a.

[0052] Here, the rotational speed of the motor unit 3 can be easily changed by varying the drive frequency using inverter control. However, if only the drive frequency is reduced without reducing the drive voltage, the motor unit 3 will burn out. Therefore, generally, the drive frequency and drive voltage are changed simultaneously. The increase or decrease in the drive frequency and the increase or decrease in the drive voltage are proportional. That is, the drive voltage decreases in proportion to the decrease in the drive frequency and increases in proportion to the increase in the drive frequency. As mentioned above, the fundamental wave component of the detection signal is generated by the drive voltage, so the signal level of the fundamental wave component is proportional to the drive frequency and drive voltage. Therefore, the signal level of the fundamental wave component increases or decreases in proportion to the increase or decrease in the drive frequency and drive voltage. Also, if the output of the motor unit 3 is the same, the drive current decreases when the drive voltage increases. That is, the drive current increases or decreases inversely proportional to the increase or decrease in the drive frequency and drive voltage. Therefore, the signal level of the fundamental wave component increases or decreases inversely proportional to the increase or decrease in the drive current. Thus, the signal level of the fundamental wave component increases or decreases in response to increases or decreases in the driving conditions (driving frequency, driving voltage value, driving current value). Pump 1 is designed to operate under rated conditions (i.e., rated frequency, rated voltage value, rated current value).

[0053] The drawing primarily referenced in this explanation is Figure 6. Detection coils C1, C3, C5, and C7 detect the amount of radial displacement of the rotor 36 (i.e., the amount of radial wear of bearings 32 and 33) by detecting the change in magnetic flux corresponding to the radial displacement of the rotor 36 due to the widening of the gaps Sr1 and Sr2. Detection coil C1 is connected to a common path Lc1 and detection coil C3. Detection coil C3 is connected to a signal processing circuit 51a via a connection part 50. Detection coils C1 and C3 are connected in series so that the difference between their detection signals is obtained (so that the detection signals cancel each other out). Detection coil C5 is connected to a common path Lc1 and detection coil C7. Detection coil C7 is connected to a signal processing circuit 51b via a connection part 50. Detection coils C5 and C7 are connected in series so that the difference between their detection signals is obtained (so that the detection signals cancel each other out). Detection coils C1, C3, C5, and C7 are examples of radial detection coils in the present invention. Detection coils C1 and C3 constitute a pair of radial detection coils and are an example of the first radial detection coil in the present invention. Detection coils C5 and C7 constitute another pair of radial detection coils and are an example of the second radial detection coil in the present invention.

[0054] In this invention, detection coils C5 and C7 may constitute one pair of radial detection coils, and detection coils C1 and C3 may constitute another pair of radial detection coils.

[0055] When the rotor 36 is displaced radially from its initial position, the signal level of the harmonic components increases as the rotor 36 approaches the detection coils C1, C3, C5, and C7, and decreases as the rotor 36 moves away from the detection coils C1, C3, C5, and C7. On the other hand, the signal level of the fundamental wave component does not increase or decrease. Therefore, in the combined signal of the detection signals from each detection coil C1 and C3 (hereinafter referred to as "combined signal (S13)"), the difference in the signal levels of the harmonic components increases in proportion to the increase in displacement. Similarly, in the combined signal of the detection signals from each detection coil C5 and C7 (hereinafter referred to as "combined signal (S57)"), the difference in the signal levels of the harmonic components increases in proportion to the increase in displacement. These differences indicate the amount of radial wear of the bearings 32 and 33, and the value of these differences is expressed as a voltage value. Therefore, for example, when there is no radial displacement of the rotor 36, the fundamental and harmonic components cancel each other out in each combined signal (S13, S57), and theoretically, the voltage value becomes "0". On the other hand, when there is radial displacement of the rotor 36, the difference in harmonic components in each combined signal (S13, S57) increases according to the amount of displacement, and the voltage value increases according to that difference.

[0056] Detection coils C2, C4, C6, and C8 detect the amount of thrust displacement of the rotor 36 (i.e., the amount of thrust wear of bearings 32 and 33) by detecting the change in magnetic flux corresponding to the thrust displacement of the rotor 36 due to the widening of the gaps Ss1 and Ss2. Detection coil C2 is connected to the common path Lc2 and detection coil C4. Detection coil C4 is connected to the signal processing circuit 51c via the connection part 50. Detection coils C2 and C4 are connected so that their detection signals are superimposed to generate a combined signal (S24). Detection coil C6 is connected to the common path Lc2 and detection coil C8. Detection coil C8 is connected to the signal processing circuit 51d via the connection part 50. Detection coils C6 and C8 are connected so that their detection signals are superimposed to generate a combined signal (S68). Detection coils C2, C4, C6, and C8 are examples of thrust detection coils in the present invention. Detection coils C2 and C4 constitute a pair of thrust detection coils and are an example of the first thrust detection coil in the present invention. Detection coils C6 and C8 constitute another pair of thrust detection coils and are an example of the second thrust detection coil in the present invention.

[0057] In this invention, detection coils C6 and C8 may constitute one set of thrust detection coils, and detection coils C2 and C4 may constitute another set of thrust detection coils.

[0058] When the rotor 36 is displaced from its initial position towards the rear, the overlap between the detection coils C2 and C4 and the front end (end ring) of the rotor 36 remains almost unchanged in the thrust direction, but the overlap between the detection coils C6 and C8 and the rear end (end ring) of the rotor 36 decreases. As a result, the signal level of the fundamental wave component of the composite signal (S24) remains almost unchanged, but the signal level of the fundamental wave component of the composite signal (S68) decreases. Similarly, when the rotor 36 is displaced from its initial position towards the front, the signal level of the fundamental wave component of the composite signal (S68) remains almost unchanged, but the signal level of the fundamental wave component of the composite signal (S24) decreases. Therefore, the amount of displacement of the rotor 36 in the thrust direction can be detected by the difference between the composite signal (S24) and the composite signal (S68). That is, the signal showing this difference (the difference signal (Sd) described later) indicates the amount of displacement of the rotor 36 in the thrust direction, i.e., the amount of wear of the bearings 32 and 33 in the thrust direction. Therefore, for example, when there is no thrust displacement of the rotor 36, the fundamental and harmonic components in the difference signal (Sd) cancel each other out, and theoretically, its voltage value is "0". On the other hand, when there is thrust displacement of the rotor 36, the signal level of the fundamental component of one side of the composite signal (e.g., composite signal (S24)) decreases according to the amount of displacement, while the signal level of the fundamental component of the other side of the composite signal (e.g., composite signal (S68)) hardly changes. As a result, the difference value (voltage value) of the fundamental component in the difference signal (Sd) increases according to the amount of displacement.

[0059] The connection section 50 is the interface to which detection coils C1 to C8 are connected. The common paths Lc1 and Lc2, detection coils C3, C4, C7, and C8, and signal processing circuits 51a to 51d are connected to the connection section 50. The common paths Lc1 and Lc2 are connected to ground via the connection section 50.

[0060] The signal processing circuits 51a to 51d are connected to the corresponding sets of detection coils C1 to C8 and perform predetermined signal processing (rectification, AC-DC conversion) on the corresponding combined signals (S13, S57, S24, S68) to convert the combined signals (S13, S57, S24, S68) from AC to DC. The signal processing circuits 51a to 51d include, for example, a filter circuit, a rectifier circuit, and an integrator circuit. The signal processing circuits 51a to 51d are connected to the A / D converter 52. The signal processing circuits 51c and 51d are connected to the A / D converter 52 and the offset processing unit 56.

[0061] The A / D converter 52 converts the analog signals input from the signal processing circuits 51a to 51d, the calculation circuit 56c (described later), and the difference absolute value conversion circuit 56d into digital signals and outputs them to the control unit 53.

[0062] The control unit 53 controls the operation of the entire device 5. The control unit 53 is composed of, for example, a processor such as a CPU (Central Processing Unit) 53a, volatile memory such as RAM (Random Access Memory) 53b which functions as a work area for the CPU 53a, and non-volatile memory such as ROM (Read Only Memory) 53c which stores various information such as this estimation program and other control programs. The control unit 53 includes a first acquisition unit 530, an estimation unit 531, a setting unit 532, a second acquisition unit 533, and a wear amount detection unit 534.

[0063] In the control unit 53, the estimation program runs and works in cooperation with the hardware resources of the device 5 to realize the estimation method described later. Furthermore, by having the processor (CPU 53a) of the control unit 53 execute the estimation program, the estimation program causes the processor to function as the first acquisition unit 530, the estimation unit 531, and the setting unit 532, and causes the processor to execute the estimation method. In addition, by having a computer execute the estimation program, the estimation program causes the computer to function as the device 5 (the estimation device).

[0064] In this invention, the estimation program may be stored in the storage unit 54. Alternatively, the estimation program may be stored in an installable or executable file format on a non-temporary storage medium (e.g., a CD (Compact Disc), DVD (Digital Versatile Disc), USB (Universal Serial Bus) memory, etc.) and provided to the device 5 via a dedicated reading medium.

[0065] The first acquisition unit 530 acquires thrust input data and radial input data. Details of the thrust input data, radial input data, and the operation of the first acquisition unit 530 will be described later. The first acquisition unit 530 is an example of an acquisition unit in the present invention.

[0066] The estimation unit 531 inputs the thrust input data acquired by the first acquisition unit 530 into the thrust learning model M11 to estimate the thrust adjustment value. The estimation unit 531 also inputs the radial input data acquired by the first acquisition unit 530 into the radial learning model M12 to estimate the radial adjustment value. Details of the operation of the thrust learning model M11, the thrust adjustment value, the radial learning model M12, the radial adjustment value, and the estimation unit 531 will be described later.

[0067] The setting unit 532 sets the thrust judgment criterion based on the thrust adjustment value estimated by the estimation unit 531. The setting unit 532 also sets the radial judgment criterion based on the radial adjustment value estimated by the estimation unit 531. Details of the thrust judgment criterion, radial judgment criterion, and the operation of the setting unit 532 will be described later.

[0068] The second acquisition unit 533 acquires the combined signal (S13, S57) and the difference signal (Sd). The specific operation of the second acquisition unit 533 will be described later.

[0069] The wear detection unit 534 detects the amount of radial wear of bearings 32 and 33 based on the voltage values ​​of the combined signals (S13, S57). The wear detection unit 534 also detects the amount of thrust wear of bearings 32 and 33 based on the voltage value of the difference signal (Sd). The specific operation of the wear detection unit 534 will be described later.

[0070] The memory unit 54 stores information necessary for the operation of the device 5 (for example, offset information, first correspondence information, second correspondence information, thrust learning model M11, and radial learning model M12). The memory unit 54 is a non-volatile memory such as EEPROM (Electrically Erasable Programmable Read-Only Memory) or flash memory.

[0071] "Offset information" is information (for example, a voltage value) indicating the offset voltage that is added to or subtracted from the composite signal (S24) in the offset processing described later. The offset information is calculated in advance under rated conditions, for example, and stored in the storage unit 54.

[0072] "Offset processing" refers to the process of adding or subtracting an offset voltage to the composite signal (S24) so ​​that, under rated conditions, the difference value between the composite signal (S24) and the composite signal (S68) (the voltage value of the difference signal (Sd)) correctly indicates the amount of displacement of the rotor 36 in the thrust direction (the amount of wear of the bearings 32 and 33 in the thrust direction). In this embodiment, the offset voltage is an added signal. Through offset processing, the magnetic center position of the rotor 36 relative to the stator 37 in the thrust direction is virtually aligned with the mechanical center position of the rotor 36 relative to the stator 37 in the thrust direction (the two center positions coincide).

[0073] The "first correspondence information" is information that shows the correspondence between the thrust position (displacement) of the rotor 36 relative to the stator 37 under rated conditions and the difference value between the composite signal (S24) and the composite signal (S68). In other words, the first correspondence information shows the correspondence between the amount of wear in the thrust direction of the bearings 32 and 33 under rated conditions and the voltage value of the difference signal (Sd). The first correspondence information includes, for example, the voltage value at the front maximum position, the voltage value at the rear maximum position, and the voltage value at the intermediate position. The voltage values ​​at the front maximum position, rear maximum position, and intermediate position are examples of thrust determination criteria in the present invention and are information included in the adjustment value in the present invention.

[0074] The "maximum front position" refers to the position where the rotor 36 has moved the furthest forward in the thrust direction when the bearing 32 has reached its maximum allowable wear. The maximum front position is set according to the design of the bearing 32, and is, for example, the position at n times the length from the intermediate position to the front position (for example, n=2). In other words, the maximum front position is set based on the front position. The ratio of the voltage value of the difference signal (Sd) to the voltage value of the maximum front position corresponds to the distance the rotor 36 has moved from the intermediate position to the front (in other words, the amount of wear of the bearing 32 in the thrust direction).

[0075] The "rear maximum position" refers to the position where the rotor 36 has moved the furthest in the rear direction when the bearing 33 has reached its maximum allowable wear in the thrust direction. The rear maximum position is set according to the design of the bearing 33, and is, for example, a position that is "n" times the length from the intermediate position to the rear position. In other words, the rear maximum position is set based on the rear position. The ratio of the voltage value of the difference signal (Sd) to the voltage value of the rear maximum position corresponds to the distance the rotor 36 has moved from the intermediate position to the rear (in other words, the amount of wear of the bearing 33 in the thrust direction).

[0076] The "Second Correspondence Information" is information that shows the correspondence between the radial position (displacement) of the rotor 36 relative to the stator 37 and the voltage value of the composite signal (S13, S57) under rated conditions. In other words, the Second Correspondence Information shows the correspondence between the amount of radial wear of bearings 32 and 33 and the voltage value under rated conditions. The Second Correspondence Information includes the voltage value at the center position and the voltage value at the radial maximum position. The voltage values ​​at the front maximum position, rear maximum position, and intermediate position are examples of radial judgment criteria in the present invention and are information included in the adjustment values ​​in the present invention.

[0077] The "maximum radial position" refers to the position where the rotor 36 has moved the most in the radial direction when bearings 32 and 33 have reached their maximum allowable wear. The maximum radial position is set according to the design of bearings 32 and 33, and is, for example, a position that is "n" times the length from the center position to the radial position. The ratio of the voltage value of the combined signal (S13, S57) to the voltage value of the maximum radial position corresponds to the distance the rotor 36 has moved radially from the center position (in other words, the amount of radial wear of bearings 32 and 33).

[0078] The display unit 55 displays the wear status of the bearings 32 and 33.

[0079] Figure 9 is a functional block diagram of the offset processing unit 56.

[0080] The offset processing unit 56 performs offset processing based on the offset information. The offset processing unit 56 comprises a D / A converter 56a, an offset voltage generation circuit 56b, an arithmetic circuit 56c, and a difference absolute value conversion circuit 56d.

[0081] The D / A converter 56a converts offset information from a digital signal to an analog signal.

[0082] The offset voltage generation circuit 56b generates an offset voltage that is added to or subtracted from the composite signal (S24) based on the offset information converted into an analog signal.

[0083] The arithmetic circuit 56c performs an offset process on the combined signal (S24) by adding or subtracting an offset voltage to the combined signal (S24), and also calculates the difference between the combined signal (S24) after the offset process and the combined signal (S68).

[0084] The difference-to-absolute value conversion circuit 56d converts the difference value calculated by the arithmetic circuit 56c into an absolute value. This signal representing the absolute value (hereinafter referred to as the "difference signal (Sd)") is converted into a digital signal by the A / D converter 52 and input to the control unit 53.

[0085] ● Adjustment of correspondence As described above, detection coils C1 to C8 generate detection signals that indicate changes in magnetic flux corresponding to changes in the position (displacement) of the rotor 36 relative to the stator 37. Based on these detection signals, the device 5 detects the amount of displacement of the rotor 36 (amount of wear of bearings 32 and 33). Therefore, in order for the detection signals to correctly indicate the amount of wear of bearings 32 and 33, it is necessary to adjust the correspondence between the amount of wear of bearings 32 and 33 (amount of displacement of the rotor 36) and the voltage value of the detection signal (first correspondence information, second correspondence information). The adjustment of the correspondence will be explained below, with the adjustment of the thrust direction as an example.

[0086] Even when the rotor 36 moves within its movable range in the thrust direction, the bearings 32 and 33 do not wear. That is, when the rotor 36 is in the front position, the bearing 32 does not wear. On the other hand, when the rotor 36 is located further forward than the front position, the bearing 32 is worn. The same applies to the rear direction. Therefore, the voltage value of the difference signal (Sd) at the front position and the rear position serves as a threshold indicating that the bearings 32 and 33 are not worn. The voltage value of the difference signal (Sd) is approximately proportional to the displacement of the rotor 36 (the amount of wear on the bearings 32 and 33). For example, when the voltage value at the front position is "V1" and the amount of wear on the bearing 32 is "Lx1", the voltage value "Vx" when the rotor 36 has moved a length "Lx(Ls1+Lx1)" from the center position is expressed as "Vx=(Lx / Ls1)×V1". Therefore, the combined signals at the front and rear positions (S24, S68) are information necessary for adjusting the correspondence in the thrust direction.

[0087] Furthermore, when the rotor 36 is in the intermediate position, the mechanical center position of the rotor 36 coincides with the mechanical center position of the stator 37 in the thrust direction. At this time, the voltage value of the difference signal (Sd) must be "0", but in reality, the position where the voltage value of the difference signal (Sd) is "0" (the magnetic center position) is offset from the mechanical center position. As mentioned above, the information necessary to make the two center positions coincide is the offset information. This offset information is calculated based on the voltage values ​​of the combined signal (S24, S68) at the intermediate position. In other words, the voltage values ​​of the combined signal (S24, S68) at the intermediate position are the information necessary to adjust the correspondence in the thrust direction.

[0088] Normally, the voltage values ​​of the composite signals (S24, S68) at each rotor position cannot be obtained unless the operator moves the rotor 36 to each position and then pumps the handling fluid in that state. In other words, obtaining these voltage values ​​requires the operator to move the rotor 36 multiple times and pump the handling fluid multiple times. To put it another way, adjusting the aforementioned correspondence requires multiple manual machine operations by the operator, which is time-consuming. On the other hand, as described later, this device 5 estimates these voltage values ​​by inputting thrust input data into the thrust learning model M11. In other words, with this device 5, machine operations and fluid pumping by the operator are not required to obtain these voltage values, and the time-consuming operation is limited to inputting the input data. Therefore, this device 5 can reduce the time required for adjustment.

[0089] ●Learning Model Next, the thrust learning model M11 and the radial learning model M12 will be described below. Figures 1 and 6 will be referenced as appropriate in the following description.

[0090] Figure 10 is a schematic diagram showing an example of information stored in the memory unit 54 (thrust learning model M11, radial learning model M12).

[0091] The "Thrust Learning Model M11" is a pre-trained machine learning algorithm (learning model) that is designed to output a thrust adjustment value when thrust input data is received. The Thrust Learning Model M11 is generated in advance by, for example, a machine learning device and stored in the memory unit 54.

[0092] The "radial learning model M12" is a pre-trained model that has been machine-trained to output a radial adjustment value when radial input data is received. The radial learning model M12 is generated in advance by, for example, a machine learning device and stored in the memory unit 54.

[0093] Machine learning is executed, for example, by causing known machine learning algorithms (e.g., neural networks having an input layer, a plurality of intermediate layers, and an output layer) to learn learning data. "Learning data" includes information that becomes input data (thrust input data, radial input data) for a machine learning algorithm (learning model), and information that becomes output data (teacher data) associated with the input data.

[0094] "Input data" is an explanatory variable in machine learning, and in the present embodiment, includes thrust input data and radial input data.

[0095] "Output data" is a target variable in machine learning, and in the present embodiment, includes thrust output data and radial output data.

[0096] "Thrust input data" is data that affects thrust output data, and includes rotor position, thrust play length, motor frame number, and rated frequency.

[0097] "Radial input data" is data that affects radial output data, and includes rotor position, motor frame number, and rated frequency.

[0098] "Thrust output data" is the voltage value of the composite signals (S24, S68) corresponding to each rotor position in the state before adjustment of the above-described correspondence relationship. Specifically, the thrust output data is the voltage value "V sff " of the composite signal (S24) and the voltage value "V srf " of the composite signal (S68) at the front position, the voltage value "V sfc " of the composite signal (S24) and the voltage value "V src " of the composite signal (S68) at the middle position, or the voltage value "V sfr " of the composite signal (S24) and the voltage value "V srr " of the composite signal (S68) at the rear position. These voltage values "V sff " "V srf " "V sfc " "V src"V sfr "V srr " is an example of a thrust adjustment value in the present invention and is information included in the adjustment value in the present invention. That is, in this embodiment, the thrust adjustment value is these voltage values ​​"V sff "V srf "V sfc "V src "V sfr "V srr This includes ".

[0099] "Radial output data" is the voltage value of the composite signal (S13, S57) corresponding to the rotor position in the state before the aforementioned correspondence adjustment. Specifically, the radial output data is the voltage value of the composite signal (S13) at the front position "V rff " and the voltage value of the combined signal (S57) "V rrf ", the voltage value of the composite signal (S13) at the intermediate position "V rfc " and the voltage value of the combined signal (S57) "V rrc " or the voltage value of the composite signal (S13) at the rear position "V rfr " and the voltage value of the combined signal (S57) "V rrr These voltage values ​​"V rff "V rrf "V rfc "V rrc "V rfr "V rrr " is an example of a radial adjustment value in the present invention and is information included in the adjustment value in the present invention. That is, in this embodiment, the radial adjustment value is these voltage values ​​"V rff "V rrf "V rfc "V rrc "V rfr "V rrr This includes ".

[0100] The "rotor position" is the position of the rotor 36 relative to the stator 37 in the thrust direction, and is one of the front position, rear position, or intermediate position. Each rotor position is assigned a numerical value, such as "0", "1", or "2". Here, the voltage value of the difference signal (Sd) decreases as the rear position moves from the intermediate position towards the rear. The voltage value of the difference signal (Sd) decreases as the front position moves from the intermediate position towards the front. Therefore, the voltage values ​​of the difference signal (Sd) at the front and rear positions change according to the thrust play length. In other words, there is a correlation between the voltage values ​​of the difference signal (Sd) at the front and rear positions and the thrust play length. To put it another way, there is a correlation between the voltage values ​​of the combined signals (S24, S68) and the thrust play length.

[0101] The "motor frame number" differs from the frame number specified in "JIS C 4210." It is a number set for each model of motor unit 3, associated not only with the size of the motor unit 3 but also with the rated output value of the motor unit 3 and the number of turns of the detection coils C1 to C8 (hereinafter simply referred to as "number of turns"). In other words, the motor frame number defines the combination of the size of the motor unit 3, the rated output value of the motor unit 3, and the number of turns. To put it another way, the motor frame number defines the combination of the size of the motor unit 3, the rated current value of the motor unit 3, the rated voltage value of the motor unit 3, and the number of turns. That is, the motor frame number includes the rated current value of the motor unit 3, the rated voltage value of the motor unit 3, and the number of turns. The voltage values ​​of the composite signals (S13, S57, S24, S68) increase or decrease in inverse proportion to the increase or decrease of the rated current value, increase or decrease in proportion to the increase or decrease of the rated voltage value, and increase or decrease in proportion to the increase or decrease of the rated frequency. In other words, there is a correlation between the voltage values ​​of the combined signals (S13, S57, S24, S68) and the rated current, rated voltage, and rated frequency. Furthermore, the voltage values ​​of the combined signals (S13, S57, S24, S68) increase or decrease according to the increase or decrease in the number of turns of the detection coils C1 to C8. That is, there is a correlation between the voltage values ​​of the combined signals (S13, S57, S24, S68) and the number of turns. To put it another way, there is a correlation between the voltage values ​​of the combined signals (S13, S57, S24, S68) and the motor frame number.

[0102] Thus, there is a correlation between thrust input data and thrust output data. As mentioned above, the training data (thrust learning data) for machine learning of the thrust learning model M11 is generated based on data (big data) accumulated during past adjustments of correspondences. That is, for example, thrust input data (rotor position, thrust play length, motor frame number, and rated frequency) for each adjustment, and thrust output data corresponding to the same thrust input data (voltage values ​​of the composite signals (S24, S68) corresponding to the rotor position) are extracted from the big data. By relating the extracted thrust input data and thrust output data to each other, a single thrust learning data is generated.

[0103] Furthermore, there is a correlation between radial input data and radial output data, except for the rotor position. In this embodiment, the training data (radial training data) for machine learning of the radial learning model M12 is generated based on the aforementioned big data. That is, for example, radial input data for each adjustment (rotor position, motor frame number, and rated frequency), and radial output data corresponding to the radial input data (voltage value of the composite signal (S13, S57) corresponding to the rotor position) are extracted from the big data. By relating the extracted radial input data and radial output data to each other, one radial training data is generated.

[0104] As shown in Figure 10, the thrust learning model M11 generated in this way can output the voltage value of a composite signal (S24, S68) corresponding to the rotor position when thrust input data (rotor position, thrust play length, motor frame number, and rated frequency) is input. In other words, the thrust learning model M11 is machine-trained to output the voltage value of a composite signal (S24, S68) corresponding to the rotor position when rotor position, thrust play length, motor frame number (rated current, rated voltage, number of turns), and rated frequency are input data. Similarly, the radial learning model M12 is machine-trained to output the voltage value of a composite signal (S13, S57) corresponding to the rotor position when rotor position, motor frame number (rated current, rated voltage, number of turns), and rated frequency are input data.

[0105] In this invention, the machine learning algorithm used for machine learning is not limited to neural networks. That is, for example, the machine learning algorithm may be a random forest, a decision tree, a support vector machine, or the like.

[0106] ● Operation of the motor bearing wear monitoring device Next, the operation of pump 1 will be explained below, focusing on the operation of device 5 before pump 1 is shipped. Figures 1 to 10 will be referenced as appropriate in the following explanation.

[0107] Figure 11 is a flowchart showing an example of the operation of the device 5.

[0108] For example, before shipping the pump 1, the device 5 performs thrust adjustment value estimation processing (ST1), radial adjustment value estimation processing (ST2), and judgment criterion setting processing (ST3). Furthermore, the device 5 performs wear amount detection processing (ST4) while the pump 1 is in operation. The thrust adjustment value estimation processing (ST1) and radial adjustment value estimation processing (ST2) are examples of this estimation method.

[0109] As mentioned above, before shipping pump 1, it is necessary to adjust the aforementioned correspondence so that the voltage values ​​of the detection coils C1 to C8 indicate the amount of wear on bearings 32 and 33.

[0110] ●Thrust adjustment value estimation process Figure 12 is a flowchart showing an example of the thrust adjustment value estimation process (ST1).

[0111] The "Thrust Adjustment Value Estimation Process (ST1)" is a process in which the device 5 (control unit 53) estimates the voltage value (thrust adjustment value) of the composite signal (S24, S68) using the thrust learning model M11.

[0112] First, the first acquisition unit 530 acquires thrust input data (rotor position, thrust play length, motor frame number, and rated frequency) for the pump 1 (ST11: acquisition step). The thrust input data for the pump 1 is acquired by the first acquisition unit 530 when, for example, the operator of the pump 1 inputs it to the device 5 using an input device (not shown).

[0113] Next, the estimation unit 531 inputs the thrust input data acquired by the first acquisition unit 530 to the thrust learning model M11 and estimates the voltage value of the composite signal (S24, S68) corresponding to the rotor position (ST12: estimation step). The estimated voltage value is associated with information indicating the rotor position and stored, for example, in the storage unit 54. Specifically, when the rotor position is the front position, the estimation unit 531 estimates the voltage value corresponding to the front position.

[0114] Next, the control unit 53 determines whether or not voltage values ​​have been estimated at all rotor positions (ST13). If voltage values ​​have not been estimated at all rotor positions, that is, if there are rotor positions for which voltage values ​​have not been estimated ("N" in ST13), the thrust adjustment value estimation process (ST1) returns to process (ST11). On the other hand, if voltage values ​​have been estimated at all rotor positions ("Y" in ST13), the thrust adjustment value estimation process (ST1) terminates.

[0115] Thus, the device 5 estimates the thrust adjustment value by inputting thrust input data into the thrust learning model M11. Therefore, with the device 5, machine operation and fluid supply by an operator are not required to acquire these voltage values, and the only labor required is inputting the input data. Consequently, the device 5 can reduce the labor required for adjustment.

[0116] ● Radial adjustment value estimation process Figure 13 is a flowchart showing an example of the radial adjustment value estimation process (ST2).

[0117] The "Radial Adjustment Value Estimation Process (ST2)" is a process in which the device 5 (control unit 53) uses the radial learning model M12 to estimate the voltage value (radial adjustment value) of the composite signal (S13, S57).

[0118] First, the first acquisition unit 530 acquires radial input data (rotor position, motor frame number, and rated frequency) from the pump 1 (ST21: acquisition step). The radial input data from the pump 1 is acquired by the first acquisition unit 530 when, for example, the operator of the pump 1 inputs it to the device 5 using an input device (not shown).

[0119] Next, the estimation unit 531 inputs the radial input data acquired by the first acquisition unit 530 to the radial learning model M12 and estimates the voltage value of the composite signal (S13, S57) corresponding to the rotor position (ST22: estimation step). The estimated voltage value is associated with information indicating the rotor position and stored, for example, in the storage unit 54. Specifically, when the rotor position is the rear position, the estimation unit 531 estimates the voltage value corresponding to the radial position at the rear position.

[0120] Next, the control unit 53 determines whether or not voltage values ​​have been estimated at all rotor positions (ST23). If voltage values ​​have not been estimated at all rotor positions, that is, if there are rotor positions for which voltage values ​​have not been estimated (ST23 "N"), the radial adjustment value estimation process (ST2) returns to process (ST21). On the other hand, if voltage values ​​have been estimated at all rotor positions (ST23 "Y"), the radial adjustment value estimation process (ST2) terminates.

[0121] Adjusting the radial alignment, like adjusting the thrust alignment, requires multiple manual machine operations by an operator, resulting in a significant amount of work. On the other hand, this device 5 estimates these voltage values ​​(radial adjustment values) by inputting radial input data into the radial learning model M12. In other words, with this device 5, obtaining these voltage values ​​does not require machine operation or fluid transfer by an operator, and the work involved is limited to inputting the input data. Therefore, this device 5 can reduce the amount of work required for adjustment.

[0122] ● Judgment criteria setting process Figure 14 is a flowchart showing an example of the judgment criterion setting process (ST3).

[0123] The "Judgment Criteria Setting Process (ST3)" is a process that sets thrust judgment criteria and radial judgment criteria based on the estimated voltage value.

[0124] First, the setting unit 532 obtains the voltage value (thrust adjustment value) of the composite signal (S24, S68) estimated by the estimation unit 531 from the storage unit 54 (ST31).

[0125] Next, the setting unit 532 calculates a thrust judgment criterion based on the acquired thrust adjustment value (ST32). Specifically, the setting unit 532 calculates the difference value of the voltage values ​​corresponding to the intermediate position. Next, the setting unit 532 adds the difference value to the voltage value of the combined signal (S24) and calculates the difference value between the voltage value of the combined signal (S24) after addition and the voltage value of the combined signal (S68). At this time, the difference value corresponding to the intermediate position is "0". Next, the setting unit 532 calculates "n" times each difference value as the thrust judgment criterion.

[0126] Next, the setting unit 532 sets the calculated thrust determination criterion as the latest thrust determination criterion for the device 5 (ST33).

[0127] Next, the setting unit 532 obtains the voltage values ​​(radial adjustment values) of the composite signals (S13, S57) estimated by the estimation unit 531 from the storage unit 54 (ST34).

[0128] Next, the setting unit 532 calculates a radial judgment criterion based on the acquired radial adjustment value (ST35). Specifically, the setting unit 532 compares the voltage value of the composite signal (S13) and the voltage value of the composite signal (S57) at a predetermined rotor position (for example, the front position). Next, the setting unit 532 selects the larger voltage value between the voltage value of the composite signal (S13) and the voltage value of the composite signal (S57). Next, the setting unit 532 calculates "n" times the selected voltage value as the radial judgment criterion.

[0129] Next, the setting unit 532 sets the calculated radial judgment criterion as the latest radial judgment criterion for the device 5 (ST36).

[0130] ● Wear amount detection process The "wear amount detection process (ST4)" is a process that detects the amount of wear of bearings 32 and 33 based on each combined signal (S13, S57, S24, S68). Since the wear amount detection process (ST4) is a known process, a detailed explanation is omitted.

[0131] Detection coils C1 to C8 constantly output detection signals while the rotor 36 is rotating. The second acquisition unit 533 acquires the combined signal (S13, S57) and the difference signal (Sd).

[0132] In the wear amount detection process (ST4), the wear amount detection unit 534 detects the radial wear amount of bearings 32 and 33 based on the second correspondence relationship information and the composite signal (S13, S57). The wear amount detection unit 534 also detects the thrust wear amount of bearings 32 and 33 based on the first correspondence relationship information and the difference signal (Sd).

[0133] ●Summary According to the embodiment described above, the device 5 comprises a storage unit 54, a first acquisition unit 530, and an estimation unit 531. The storage unit 54 stores a thrust learning model M11. The thrust learning model M11 is machine-trained to output a thrust adjustment value when thrust input data is input. The thrust input data includes at least the rotor position, thrust play length, rated current value, and number of turns. The rotor position is one of the front position, rear position, or intermediate position. The thrust adjustment value is the voltage value of the combined signal (S24, S68) when the rotor 36 is located in the front position, rear position, and intermediate position, respectively. With this configuration, manual mechanical operation for adjusting the correspondence relationship (first correspondence relationship information) is unnecessary. As a result, the adjustment man-hours for the pump 1 are significantly reduced.

[0134] Furthermore, according to the embodiment described above, the thrust input data includes the rated voltage value and rated frequency of the motor unit 3. With this configuration, the accuracy of the thrust adjustment value estimation by the thrust learning model M11 is improved.

[0135] Furthermore, according to the embodiment described above, the thrust input data includes the motor frame number instead of the number of turns and the rated voltage value. This configuration reduces the amount of data required for machine learning. In addition, it makes it easier to generate thrust learning data.

[0136] Furthermore, according to the embodiments described above, the memory unit 54 stores the radial learning model M12. The radial learning model M12 is machine-trained to output a radial adjustment value when radial input data is input. The radial input data includes at least the rated current value and the number of turns. The radial adjustment value is the voltage value of the composite signal (S13, S57) when the rotor 36 is in the radial position. With this configuration, fluid supply for adjusting the correspondence relationship (second correspondence relationship information) is unnecessary.

[0137] ●Estimation device● Next, a second embodiment of the estimation device will be described below, using the case where the estimation device is separate from the pump as an example. In the following description, elements that are the same as in the first embodiment and elements that have common functions are denoted by the same reference numerals as in the first embodiment for the sake of explanation, and their descriptions will be omitted. Figures 1 to 5 will be referred to as appropriate in the following description.

[0138] ● Configuration of the estimation device Figure 15 is a functional block diagram of the estimation device, showing another embodiment (second embodiment) of the estimation device.

[0139] The estimation device 6 estimates the adjustment values ​​of the device 5. The estimation device 6 comprises a communication unit 60, an operation unit 61, a display unit 62, a control unit 63, and a storage unit 64. The estimation device 6 is implemented, for example, by a PC (Personal Computer). In the second embodiment, the estimation device 6 estimates the thrust judgment criterion as the thrust adjustment value and the radial judgment criterion as the radial adjustment value. That is, the adjustment value includes the thrust judgment criterion and the radial judgment criterion.

[0140] The communication unit 60 is, for example, a communication module or a communication interface. The communication unit 60 is connected to, for example, this device 5 via a telecommunications line.

[0141] The operation unit 61 is a device operated by the user of the estimation device 6 (for example, an adjustment worker) (for example, by inputting information such as input data). The operation unit 61 is, for example, a keyboard, mouse, or touch panel.

[0142] The display unit 62 is a device that displays information input screens from the operation unit 61 and information output by the control unit 63. The display unit 62 is, for example, a monitor or display.

[0143] The control unit 63 controls the overall operation of the estimation device 6. The control unit 63 is composed of, for example, a processor such as a CPU 63a, volatile memory such as a RAM 63b that functions as a work area for the CPU 53a, and non-volatile memory such as a ROM 63c that stores various information such as the estimation program. The control unit 63 includes an acquisition unit 630 and an estimation unit 631.

[0144] In the control unit 63, the estimation program runs and works in cooperation with the hardware resources of the estimation device 6 to realize the estimation method described later. Furthermore, by having the processor (CPU 63a) of the control unit 63 execute the estimation program, the estimation program causes the processor to function as an acquisition unit 630 and an estimation unit 631, and causes the processor to execute the estimation method. In addition, by having a computer execute the estimation program, the estimation program causes the computer to function as the estimation device 6.

[0145] In this invention, the estimation program may be stored in the storage unit 64. Alternatively, the estimation program may be stored in an installable or executable file format on a non-temporary storage medium (e.g., a CD, DVD, USB memory, etc.) and provided to the estimation device 6 via a dedicated reading medium.

[0146] The acquisition unit 630 acquires thrust input data and radial input data. Details of the operation of the acquisition unit 630 will be described later.

[0147] The estimation unit 631 inputs the thrust input data acquired by the acquisition unit 630 into the thrust learning model M21 to estimate the thrust adjustment value (thrust judgment criterion). The estimation unit 631 also inputs the radial input data acquired by the acquisition unit 630 into the radial learning model M22 to estimate the radial adjustment value (radial judgment criterion). Details of the operation of the estimation unit 631 will be described later.

[0148] The memory unit 64 stores information necessary for the operation of the estimation device 6 (for example, the thrust learning model M21 and the radial learning model M22). The memory unit 64 is a non-volatile memory such as an EEPROM or flash memory.

[0149] ●Learning Model Next, the thrust learning model M21 and the radial learning model M22 will be described below. Figure 15 will be referred to as appropriate in the following description.

[0150] Figure 16 is a schematic diagram showing an example of information stored in the memory unit 64 (thrust learning model M21, radial learning model M22).

[0151] The "Thrust Learning Model M21" is a pre-trained model that has been machine-trained to output a thrust judgment criterion when thrust input data is received. The Thrust Learning Model M21 is generated in advance by, for example, a machine learning device and stored in the memory unit 64.

[0152] The "radial learning model M22" is a pre-trained model that has been machine-trained to output a radial judgment criterion when radial input data is received. The radial learning model M22 is generated in advance by, for example, a machine learning device and stored in the memory unit 64.

[0153] "Thrust output data" is the thrust judgment criterion, and specifically, it is the difference between the voltage value of the combined signal (S24) after offset processing at the maximum front position and the voltage value of the combined signal (S68) (i.e., the voltage value of the difference signal (Sd)) "V df ", the same difference value at the rear maximum position "V dr ", and the same difference value "V" at the intermediate position dc In the first embodiment, these difference values ​​(thrust judgment criteria) are calculated based on the voltage values ​​of the composite signals (S24, S68) at each rotor position. There is a correlation between these difference values ​​and the thrust input data. Therefore, there is also a correlation between the thrust judgment criteria and the thrust input data.

[0154] Thus, there is a correlation between thrust input data and thrust output data. The training data (thrust learning data) for machine learning of the thrust learning model M21 is generated based on the aforementioned big data. That is, for example, thrust input data for each adjustment (rotor position, thrust play length, motor frame number, and rated frequency) and thrust output data (thrust judgment criteria) corresponding to the thrust input data are extracted from the big data. By relating the extracted thrust input data and thrust output data to each other, a single thrust learning data is generated.

[0155] "Radial output data" is the radial judgment criterion, specifically the voltage value of the composite signal (S13, S57) at the maximum radial position "V d As mentioned above, the radial judgment criteria are calculated based on the voltage values ​​of the composite signals (S13, S57) at the radial positions. There is a correlation between these voltage values ​​and the radial input data. Therefore, there is also a correlation between the radial judgment criteria and the radial input data.

[0156] Furthermore, there is a correlation between radial input data and radial output data. The training data (radial training data) for machine learning of the radial learning model M22 is generated based on the aforementioned big data. That is, for example, radial input data for each adjustment (rotor position, motor frame number, and rated frequency), and radial output data (radial judgment criteria) corresponding to the radial input data are extracted from the big data. By relating the extracted radial input data and radial output data to each other, a single radial training data is generated.

[0157] As shown in Figure 16, the thrust learning model M21 generated in this way can output thrust criteria when thrust input data (rotor position, thrust play length, motor frame number, and rated frequency) is input. In other words, the thrust learning model M21 is machine-trained to output thrust criteria when rotor position, thrust play length, motor frame number (rated current, rated voltage, number of turns), and rated frequency are input. Similarly, the radial learning model M22 is machine-trained to output radial criteria when rotor position, motor frame number (rated current, rated voltage, number of turns), and rated frequency are input.

[0158] ● Operation of the estimation device Next, the operation of the estimation device 6 will be explained below.

[0159] Figure 17 is a flowchart showing an example of the operation of the estimation device 6.

[0160] The estimation device 6 performs, for example, thrust adjustment value estimation processing (ST5) and radial adjustment value estimation processing (ST6) before shipping the pump 1. The thrust adjustment value estimation processing (ST5) and radial adjustment value estimation processing (ST6) are examples of this method.

[0161] ●Thrust adjustment value estimation process Figure 18 is a flowchart showing an example of the thrust adjustment value estimation process (ST5).

[0162] The "thrust adjustment value estimation process (ST5)" is a process in which the estimation device 6 (control unit 63) uses the thrust learning model M21 to estimate the thrust judgment criterion (thrust adjustment value).

[0163] First, the acquisition unit 630 acquires thrust input data (rotor position, thrust play length, motor frame number, and rated frequency) for the pump 1 (ST51: acquisition step). The thrust input data for the pump 1 is acquired by the acquisition unit 630 when, for example, the operator of the pump 1 inputs it to the estimation device 6 using the operation unit 61.

[0164] Next, the estimation unit 631 inputs the thrust input data acquired by the acquisition unit 630 to the thrust learning model M21 and estimates the voltage value corresponding to the thrust judgment criterion position (front maximum position, rear maximum position, or intermediate position) (ST52: estimation step). The estimated voltage value is associated with information indicating the thrust judgment criterion position and stored, for example, in the storage unit 64. Specifically, when the rotor position is the front position, the estimation unit 631 estimates the voltage value corresponding to the front maximum position.

[0165] Next, the control unit 63 determines whether or not voltage values ​​have been estimated at all rotor positions (ST53). If voltage values ​​have not been estimated at all rotor positions (ST53's "N"), the thrust adjustment value estimation process (ST5) returns to process (ST51). On the other hand, if voltage values ​​have been estimated at all rotor positions (ST53's "Y"), the thrust adjustment value estimation process (ST5) terminates. The estimated voltage values ​​are set in the device 5 as thrust determination criteria.

[0166] ● Radial adjustment value estimation process Figure 19 is a flowchart showing an example of the radial adjustment value estimation process (ST6).

[0167] The "Radial Adjustment Value Estimation Process (ST6)" is a process in which the estimation device 6 (control unit 63) uses the radial learning model M22 to estimate the radial judgment criterion (radial adjustment value).

[0168] First, the acquisition unit 630 acquires radial input data (rotor position, motor frame number, and rated frequency) from the pump 1 (ST61: acquisition step). The radial input data from the pump 1 is acquired by the acquisition unit 630 when, for example, the operator of the pump 1 inputs it to the estimation device 6 using the operation unit 61.

[0169] Next, the estimation unit 631 inputs the radial input data acquired by the acquisition unit 630 into the radial learning model M22 to estimate the voltage value at the radial maximum position (ST62: estimation step). The estimated voltage value is associated with information indicating the radial maximum position and stored, for example, in the storage unit 64. The estimated voltage value is set in the device 5 as a radial determination criterion.

[0170] ●Summary According to the embodiment described above, the estimation device 6 comprises a storage unit 64, an acquisition unit 630, and an estimation unit 631. The storage unit 64 stores a thrust learning model M21. The thrust learning model M21 is machine-trained to output a thrust adjustment value when thrust input data is input. The thrust input data includes at least the rotor position, thrust play length, rated current value, and number of turns. The rotor position is one of the front position, rear position, or intermediate position. The thrust adjustment value is the voltage value of the difference signal (Sd) when the rotor 36 is in the front maximum position, rear maximum position, and intermediate position, respectively. With this configuration, manual mechanical operation for adjusting the correspondence relationship (first correspondence relationship information) is unnecessary. As a result, the adjustment man-hours for the pump 1 are significantly reduced.

[0171] Furthermore, according to the embodiment described above, the thrust input data includes the rated voltage value and rated frequency of the motor unit 3. With this configuration, the accuracy of the thrust adjustment value estimation by the thrust learning model M21 is improved.

[0172] Furthermore, according to the embodiment described above, the thrust input data includes the motor frame number instead of the number of turns and the rated voltage value. This configuration reduces the amount of data required for machine learning. In addition, it makes it easier to generate thrust learning data.

[0173] Furthermore, according to the embodiments described above, the memory unit 64 stores the radial learning model M22. The radial learning model M22 is machine-trained to output a radial adjustment value when radial input data is input. The radial input data includes at least the rated current value and the number of turns. The radial adjustment value is the voltage value of the combined signal (S13, S57) when the rotor 36 is in the radial maximum position, with the larger value being the voltage value. With this configuration, fluid supply for adjusting the correspondence relationship (second correspondence relationship information) is unnecessary.

[0174] ●Other Embodiments● In each embodiment, the thrust input data for machine learning only needs to include at least the rated current value, thrust play length, number of turns, and rotor position of the motor unit 3, and does not need to include the rated voltage value and / or rated frequency. In this case, when the thrust learning model M11 receives the rated current value, thrust play length, number of turns, and rotor position as input, it outputs the voltage value of the composite signal (S24, S68) corresponding to that rotor position as the thrust adjustment value. Similarly, when the thrust learning model M21 receives the rated current value, thrust play length, number of turns, and rotor position as input, it outputs the difference value of the position corresponding to that rotor position as the thrust adjustment value. In this configuration, the estimation accuracy of the thrust adjustment value may be lower than in each embodiment. However, an estimation accuracy sufficient to monitor the wear state of bearings 32 and 33 by this device 5 can be obtained.

[0175] Furthermore, in each embodiment, the value of the composite signal (S13) at each rotor position is approximately the same, and the value of the composite signal (S57) at each rotor position is also approximately the same. Therefore, the radial input data only needs to include the rated current value and number of turns of the motor unit 3 at at least one rotor position, and does not need to include the rotor position. In this case, when the rated current value and number of turns are input to the radial learning model M12, it outputs the voltage value of the composite signal (S13) or composite signal (S57) as the radial adjustment value. Also, when the rated current value and number of turns are input to the radial learning model M22, it outputs the voltage value at the radial maximum position as the radial adjustment value. In this configuration, the estimation accuracy of the radial adjustment value may be lower than in each embodiment. However, an estimation accuracy sufficient to monitor the wear state of the bearings 32 and 33 by this device 5 can be obtained.

[0176] Furthermore, in the first embodiment, the memory unit 54 may store the thrust learning model M21. In this case, the estimation unit 531 estimates the difference value as the thrust adjustment value.

[0177] Furthermore, in the second embodiment, the memory unit 64 may store the thrust learning model M11. In this case, the estimation unit 631 estimates the voltage values ​​of the composite signals (S24, S68) as thrust adjustment values. The control unit 63 may function as a setting unit 532.

[0178] Furthermore, in the first embodiment, the storage unit 54 may also store the radial learning model M22. In this case, the estimation unit 631 estimates the voltage value at the radial maximum position as the radial adjustment value.

[0179] Furthermore, in the second embodiment, the storage unit 64 may store the radial learning model M12. In this case, the estimation unit 631 estimates the voltage values ​​of the composite signals (S13, S57) as radial adjustment values. The control unit 63 may function as a setting unit 532.

[0180] Furthermore, in the second embodiment, the thrust learning model M21 may output not only the difference values ​​but also offset information as thrust adjustment values. In this case, the thrust learning data of the thrust learning model M21 also includes offset information.

[0181] Furthermore, in the first embodiment, when the drive frequency (rated frequency) is changed by inverter control, the device 5 may perform processing (ST1 to ST3) to update the thrust judgment criterion and radial judgment criterion. When the drive frequency is changed by the inverter, the voltage values ​​of the detection signals of the detection coils C1 to C8 change, leading to technical problems such as false detection due to mismatches in the correspondence. Therefore, manual adjustment of the correspondence is required each time the drive frequency is changed. In this configuration, even if the drive frequency is changed by the inverter, the device 5 can adjust the correspondence without human mechanical operation.

[0182] Furthermore, in the second embodiment, the device 5 does not need to include the first acquisition unit 530, the estimation unit 531, and the setting unit 532.

[0183] ●Embodiments of the present invention● Next, embodiments of the present invention as understood from the embodiments described above will be described below, with reference to the terms and reference numerals described in each embodiment.

[0184] A first embodiment of the present invention estimates the adjustment value of a motor bearing wear monitoring device (e.g., motor bearing wear monitoring device 5) that monitors the wear state of bearings (e.g., bearings 32, 33) supporting the rotation shaft (e.g., rotation shaft 31) of a canned motor pump (e.g., canned motor pump 1) based on the detection signals of each of a plurality of detection coils (e.g., detection coils C1 to C8) that detect changes in magnetic flux corresponding to changes in the position of the rotor (e.g., stator 37) of the motor (e.g., motor unit 3). Estimation device (e.g., control unit 53, estimation device 6), wherein each of the plurality of detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, and the plurality of detection coils include a plurality of thrust detection coils (e.g., detection coils C2, C4, C6, C8) that detect the change in magnetic flux in the thrust direction of the rotation axis, and the plurality of thrust detection coils include a pair of first thrust detection coils (e.g., detection coils C2, C4) and another pair of second thrust detection coils (e.g., detection coil C6,C8) and, in the thrust direction, the first thrust detection coil is located at one end of the stator, the second thrust detection coil is located at the other end of the stator, the adjustment value includes a thrust adjustment value in the thrust direction, the thrust adjustment value includes a thrust determination criterion for the wear state in the thrust direction, which can be calculated based on the voltage value of a first thrust composite signal (e.g., composite signal (S24)) obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is located at a first position (e.g., front position), a second position (e.g., rear position), and a third position (e.g., intermediate position) relative to the stator, the voltage value of a second thrust composite signal (e.g., composite signal (S68)) obtained by combining the detection signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, the first position is the bearing in the thrust direction The first position is the position in which the rotor is moved to the maximum extent in one direction relative to the stator within the mechanically movable range of the rotating shaft, the second position is the position in which the rotor is moved to the maximum extent in the other direction relative to the stator within the movable range, and the third position is an intermediate position between the first and second positions, and the thrust input includes the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction. The estimation device comprises: a storage unit (e.g., storage units 54, 64) that stores a trained thrust learning model (e.g., thrust learning models M11, M21) that has been machine-learned to output the thrust adjustment value when data is input; an acquisition unit (e.g., first acquisition unit 530, acquisition unit 630) that acquires the thrust input data; and an estimation unit (e.g., estimation units 531, 631) that inputs the thrust input data acquired by the acquisition unit into the thrust learning model to estimate the thrust adjustment value. This configuration significantly reduces the amount of work required for adjustments.

[0185] A second embodiment of the present invention is an estimation device in which, in the first embodiment, the thrust input data includes the rated voltage value of the motor and / or the rated frequency of the motor. This configuration improves the accuracy of the thrust adjustment value estimation by the thrust learning model.

[0186] A third embodiment of the present invention is an estimation device in which, in the first embodiment, the thrust input data includes, instead of the number of turns, a motor frame number that defines a combination of the motor output, the motor size, and the number of turns. This configuration reduces the amount of data required for machine learning and makes it easier to generate thrust training data.

[0187] A fourth embodiment of the present invention, in the first embodiment, comprises a plurality of detection coils, each comprising a plurality of radial detection coils (e.g., detection coils C1, C3, C5, C7) for detecting the change in magnetic flux in the radial direction of the rotating shaft, wherein the plurality of radial detection coils comprises a pair of first radial detection coils (e.g., detection coils C1, C3) and another pair of second radial detection coils (e.g., detection coils C5, C7), wherein in the thrust direction, the first radial detection coils are located at the one-direction end of the stator, and the second radial detection coils are located at the other-direction end of the stator, the bearings are sliding bearings lubricated by the fluid handled by the canned motor pump, the adjustment value comprises a radial adjustment value in the radial direction, and the radial adjustment value is a first radial combined signal (e.g., combined) obtained by combining the detection signals output from each of the first radial detection coils when the rotor relative to the stator is in a fourth position (e.g., radial position). The estimation device includes a radial determination criterion for the wear state in the radial direction, which is generated based on the voltage value of the signal (S13) and the voltage value of a second radial composite signal (e.g., composite signal (S57)) obtained by combining the detection signals output from each of the second radial detection coils, or the voltage value of the first radial composite signal and the voltage value of the second radial composite signal, wherein the fourth position is the position in which the rotor has moved the maximum in the radial direction relative to the stator within the mechanically movable range of the rotation shaft relative to the bearing in the radial direction, the storage unit stores a trained radial learning model (e.g., radial learning models M12, M22) which has been machine-learned to output the radial adjustment value when radial input data including the rated current value and the number of turns is input, the acquisition unit acquires the radial input data, and the estimation unit inputs the radial input data acquired by the acquisition unit to the radial learning model to estimate the radial adjustment value. With this configuration, liquid delivery for adjusting the correspondence (second correspondence information) becomes unnecessary.

[0188] A fifth embodiment of the present invention is an estimation program (e.g., this estimation program) that causes a computer to function as the estimation device described in the first embodiment. This configuration significantly reduces the amount of work required for adjustments.

[0189] A sixth embodiment of the present invention provides an estimation device (e.g., control unit 53, estimation device 6) that estimates adjustment values ​​for a motor bearing wear monitoring device (e.g., motor bearing wear monitoring device 5) that monitors the wear state of bearings (e.g., bearings 32, 33) supporting the rotation shaft (e.g., rotation shaft 31) of a canned motor pump (e.g., canned motor pump 1), based on the detection signals of each of a plurality of detection coils (e.g., detection coils C1 to C8) that detect changes in magnetic flux corresponding to changes in the position of the rotor (e.g., stator 37) of the motor (e.g., motor unit 3), and provides an estimation device (e.g., control unit 53, estimation device 6) that estimates adjustment values ​​for a motor bearing wear monitoring device (e.g., motor bearing wear monitoring device 5) that monitors the wear state of bearings (e.g., bearings 32, 33) supporting the rotation shaft (e.g., rotation shaft 31) of the canned motor pump (e.g., canned motor pump 1). A method for estimating the adjustment value (e.g., thrust adjustment value estimation process (ST1, ST5)) to be performed, wherein each of the plurality of detection coils is attached to the stator and outputs a detection signal indicating the change in magnetic flux, and the plurality of detection coils include a plurality of thrust detection coils (e.g., detection coils C2, C4, C6, C8) that detect the change in magnetic flux in the thrust direction of the rotation axis, and the plurality of thrust detection coils include a pair of first thrust detection coils (e.g., detection coils C2, C4) and another pair of second thrust detection coils (e.g., detection coil C6,C8) and, in the thrust direction, the first thrust detection coil is located at one end of the stator, and the second thrust detection coil is located at the other end of the stator, and the adjustment value includes a thrust adjustment value in the thrust direction, the thrust adjustment value being the output from each of the first thrust detection coils when the rotor is located at a first position (e.g., front position), a second position (e.g., rear position), and a third position (e.g., intermediate position) relative to the stator. The system includes a thrust determination criterion for the wear state in the thrust direction, which is generated based on the voltage value of a first thrust composite signal (e.g., composite signal (S24)) obtained by combining the detection signals, the voltage value of a second thrust composite signal (e.g., composite signal (S68)) obtained by combining the detection signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, wherein the first position is within the mechanically movable range of the rotating shaft relative to the bearing in the thrust direction. The first position is the position where the rotor has moved the most in one direction relative to the stator, the second position is the position where the rotor has moved the most in the other direction relative to the stator within the movable range, and the third position is an intermediate position between the first and second positions. The estimation device outputs the thrust adjustment value when thrust input data is input, which includes the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction. The estimation method comprises a storage unit (e.g., storage units 54, 64) that stores a trained thrust learning model (e.g., thrust learning models M11, M21) that has been machine-learned to perform the following actions: the estimation device includes an acquisition step (e.g., acquisition step (ST11, ST51)) for acquiring the thrust input data, and an estimation step (e.g., estimation step (ST12, ST52)) for inputting the thrust input data acquired in the acquisition step into the thrust learning model to estimate the thrust adjustment value. This configuration significantly reduces the amount of work required for adjustments.

[0190] A seventh embodiment of the present invention is a motor bearing wear monitoring device (e.g., motor bearing wear monitoring device 5) that monitors the wear state of bearings (e.g., bearings 32, 33) supporting the rotating shaft (e.g., rotating shaft 31) of a canned motor pump (e.g., canned motor pump 1) based on the detection signals of each of the plurality of detection coils (e.g., detection coils C1 to C8) that detect changes in magnetic flux corresponding to changes in the position of the rotor (e.g., rotor 36) relative to the stator (e.g., stator 37) of the motor (e.g., motor unit 3), wherein each of the plurality of detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, and the plurality of detection coils include a plurality of thrust detection coils (e.g., detection coils C2, C4, C6, C8) that detect the change in magnetic flux in the thrust direction of the rotating shaft, and the plurality of thrust detection coils include a pair of first thrust detection coils (e.g., detection coils C2, C4) and another pair of second thrust detection coils (e.g., detection coil C6,The motor includes C8), wherein in the thrust direction, the first thrust detection coil is positioned at one end of the stator, and the second thrust detection coil is positioned at the other end of the stator, and the motor includes a storage unit (e.g., storage unit 54) that stores a trained thrust learning model (e.g., thrust learning model M11) that has been machine-trained to output a thrust adjustment value when thrust input data is input, which includes the rated current value of the motor, the length of the mechanically movable range of the rotating shaft relative to the bearing in the thrust direction, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction; an acquisition unit (e.g., first acquisition unit 530) that acquires the thrust input data; and an estimation unit (e.g., estimation unit 531) that inputs the thrust input data acquired by the acquisition unit into the thrust learning model to estimate the thrust adjustment value, wherein the thrust adjustment value is determined when the rotor is in a first position relative to the stator (e.g., The motor bearing wear monitoring device includes a thrust determination criterion for the wear state in the thrust direction, which can be calculated based on the voltage value of a first thrust composite signal (e.g., composite signal (S24)) obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is located at the front position, a second position (e.g., rear position), and a third position (e.g., intermediate position), the voltage value of a second thrust composite signal (e.g., composite signal (S68)) obtained by combining the detection signals output from each of the second thrust detection coils, or the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal, wherein the first position is the position in which the rotor has moved the maximum distance in one direction relative to the stator within the movable range, the second position is the position in which the rotor has moved the maximum distance in the other direction relative to the stator within the movable range, and the third position is an intermediate position between the first and second positions. This configuration significantly reduces the amount of work required for adjustments.

[0191] An eighth embodiment of the present invention is a canned motor pump (e.g., canned motor pump 1) comprising a rotor (e.g., rotor 36), a stator (e.g., stator 37) that rotates the rotor, a rotating shaft (e.g., rotating shaft 31) to which the rotation of the rotor is transmitted, bearings (e.g., bearings 32, 33) that rotatably support the rotating shaft, a plurality of detection coils (e.g., detection coils C1 to C8) attached to the stator that detect changes in magnetic flux corresponding to changes in the position of the rotor relative to the stator, and a motor bearing wear monitoring device (e.g., motor bearing wear monitoring device 5) that monitors the wear state of the bearings based on the output signals of the detection coils, wherein the motor bearing wear monitoring device is the motor bearing wear monitoring device described in the seventh embodiment. This configuration significantly reduces the amount of work required for adjustments. [Explanation of symbols]

[0192] 1. Canned motor pump 3. Motor section (motor) 5. Motor bearing wear monitoring device 53 Control Unit (Estimation Device) 530 1st Acquisition Department (Acquisition Department) 531 Estimation Department 54 Memory section 6 Estimation device 630 Acquisition Department 631 Estimation Department 64 Memory section C1, C3 detection coils (radial detection coil, first radial detection coil) C5, C7 detection coils (radial detection coil, second radial detection coil) C2, C4 detection coils (thrust detection coil, first thrust detection coil) C6, C8 detection coils (thrust detection coil, second thrust detection coil) M11 Thrust Learning Model M12 Radial Learning Model M21 Thrust Learning Model M22 Radial Learning Model

Claims

1. An estimation device for estimating adjustment values ​​for a motor bearing wear monitoring device that monitors the wear state of the bearing supporting the rotation axis of a canned motor pump, based on the detection signals of each of a plurality of detection coils that detect magnetic flux changes corresponding to changes in the position of the rotor relative to the stator of the motor, Each of the multiple detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, Multiple detection coils, A plurality of thrust detection coils for detecting the change in magnetic flux in the thrust direction of the rotation axis, Includes, Multiple thrust detection coils, A set of first thrust detection coils, The other set consists of a second thrust detection coil and Includes, In the thrust direction, The first thrust detection coil is positioned at one end of the stator, The second thrust detection coil is positioned at the other end of the stator, The aforementioned adjustment value, The thrust adjustment value in the thrust direction, Includes, The thrust adjustment value is, The voltage value of the first thrust composite signal obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is in the first, second, and third positions relative to the stator, and the voltage value of the second thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, or A thrust determination criterion for the wear state in the thrust direction, which can be calculated based on the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal. Includes, The first position is the position in which the rotor moves to the maximum extent possible in one direction relative to the stator, within the mechanically movable range of the rotation axis relative to the bearing in the thrust direction. The second position is the position in which the rotor moves to the maximum extent possible relative to the stator in the other direction within the movable range. The third position is an intermediate position between the first position and the second position. A storage unit stores a trained thrust learning model that has been machine-learned to output the thrust adjustment value when thrust input data is input, including the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction. The acquisition unit acquires the aforementioned thrust input data, An estimation unit inputs the thrust input data acquired by the acquisition unit into the thrust learning model to estimate the thrust adjustment value, Having, Estimation device.

2. The thrust input data is, The rated voltage value of the motor, and / or the rated frequency of the motor, including, The estimation device according to claim 1.

3. The thrust input data is, Instead of the aforementioned number of turns, a motor frame number is determined by a combination of the motor's output, the motor's size, and the aforementioned number of turns. including, The estimation device according to claim 1.

4. Multiple detection coils, Multiple radial detection coils for detecting the change in magnetic flux in the radial direction of the rotation axis, Includes, Multiple radial detection coils, A set of first radial detection coils, The other pair consists of a second radial detection coil, Includes, In the thrust direction, The first radial detection coil is positioned at the unidirectional end of the stator, The second radial detection coil is positioned at the other end of the stator, The bearing is a sliding bearing lubricated by the fluid handled by the canned motor pump. The aforementioned adjustment value, The radial adjustment value in the radial direction, Includes, The radial adjustment value is, When the rotor is in the fourth position relative to the stator, the voltage value of the first radial combined signal obtained by combining the detection signals output from each of the first radial detection coils, and the voltage value of the second radial combined signal obtained by combining the detection signals output from each of the second radial detection coils, or A radial determination criterion for the wear state in the radial direction, which is generated based on the voltage value of the first radial composite signal and the voltage value of the second radial composite signal. Includes, The fourth position is the position in which the rotor moves to its maximum extent in the radial direction relative to the stator, within the mechanically movable range of the rotation axis relative to the bearing in the radial direction. The memory unit stores a trained radial learning model that has been machine-learned to output the radial adjustment value when radial input data including the rated current value and the number of turns is input. The acquisition unit acquires the radial input data, The estimation unit inputs the radial input data acquired by the acquisition unit into the radial learning model to estimate the radial adjustment value. The estimation device according to claim 1.

5. The computer is made to function as the estimation device described in claim 1. Estimation program.

6. A method for estimating an adjustment value for a motor bearing wear monitoring device that monitors the wear state of the bearing supporting the rotation axis of a canned motor pump, based on the detection signals of each of a plurality of detection coils that detect magnetic flux changes corresponding to changes in the position of the rotor relative to the stator of the motor, wherein the device is used to estimate the adjustment value for the motor bearing wear monitoring device that monitors the wear state of the bearing supporting the rotation axis of the rotor, Each of the multiple detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, Multiple detection coils, A plurality of thrust detection coils for detecting the change in magnetic flux in the thrust direction of the rotation axis, Includes, Multiple thrust detection coils, A set of first thrust detection coils, The other set consists of a second thrust detection coil and Includes, In the thrust direction, The first thrust detection coil is positioned at one end of the stator, The second thrust detection coil is positioned at the other end of the stator, The aforementioned adjustment value, The thrust adjustment value in the thrust direction, Includes, The thrust adjustment value is, The voltage value of the first thrust composite signal obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is in the first, second, and third positions relative to the stator, and the voltage value of the second thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, or A thrust determination criterion for the wear state in the thrust direction, which is generated based on the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal. Includes, The first position is the position in which the rotor moves to the maximum extent possible in one direction relative to the stator, within the mechanically movable range of the rotation axis relative to the bearing in the thrust direction. The second position is the position in which the rotor moves to the maximum extent possible relative to the stator in the other direction within the movable range. The third position is an intermediate position between the first position and the second position. The estimation device is, A storage unit stores a trained thrust learning model that has been machine-learned to output the thrust adjustment value when thrust input data is input, including the rated current value of the motor, the length of the movable range, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction. Equipped with, The estimation device includes an acquisition step of acquiring the thrust input data, The estimation device inputs the thrust input data acquired in the acquisition step into the thrust learning model to estimate the thrust adjustment value, including, Estimation method.

7. A motor bearing wear monitoring device that monitors the wear state of the bearing supporting the rotation axis of a canned motor pump based on the detection signals from each of a plurality of detection coils that detect magnetic flux changes corresponding to changes in the position of the rotor relative to the stator of the motor, Each of the multiple detection coils is attached to the stator and outputs the detection signal indicating the change in magnetic flux, Multiple detection coils, A plurality of thrust detection coils for detecting the change in magnetic flux in the thrust direction of the rotation axis, Includes, Multiple thrust detection coils, A set of first thrust detection coils, The other set consists of a second thrust detection coil and Includes, In the thrust direction, The first thrust detection coil is positioned at one end of the stator, The second thrust detection coil is positioned at the other end of the stator, A storage unit stores a trained thrust learning model that has been machine-learned to output a thrust adjustment value when thrust input data is input, which includes the rated current value of the motor, the length of the mechanically movable range of the rotating shaft relative to the bearing in the thrust direction, the number of turns of the detection coil, and the position of the rotor relative to the stator in the thrust direction. The acquisition unit acquires the aforementioned thrust input data, An estimation unit inputs the thrust input data acquired by the acquisition unit into the thrust learning model to estimate the thrust adjustment value, It has, The thrust adjustment value is, The voltage value of the first thrust composite signal obtained by combining the detection signals output from each of the first thrust detection coils when the rotor is in the first, second, and third positions relative to the stator, and the voltage value of the second thrust composite signal obtained by combining the detection signals output from each of the second thrust detection coils, or A thrust determination criterion for the wear state in the thrust direction, which is generated based on the voltage value of the first thrust composite signal and the voltage value of the second thrust composite signal. Includes, The first position is the position in which the rotor moves to the maximum extent possible in one direction relative to the stator within the movable range. The second position is the position in which the rotor moves to the maximum extent possible relative to the stator in the other direction within the movable range. The third position is an intermediate position between the first position and the second position. Motor bearing wear monitoring device.

8. Rotor and, A stator that rotates the rotor, The rotating shaft through which the rotation of the rotor is transmitted, A bearing that rotatably supports the aforementioned rotating shaft, A plurality of detection coils attached to the stator are used to detect changes in magnetic flux corresponding to changes in the rotor's position relative to the stator, A motor bearing wear monitoring device that monitors the wear state of the bearing based on the output signal of the detection coil, It has, The motor bearing wear monitoring device is the motor bearing wear monitoring device described in claim 7. Canned motor pump.

Images