JPH07311082A - Failure diagnostic device of rotating equipment - Google Patents

Abstract

PURPOSE:To simplify the signal processing for failure diagnosis by setting rotary angle space signals as signals of a common constant period even if the number of revolutions of a rotating shaft fluctuates with time or if the number of revolutions of the rotating shaft varies. CONSTITUTION:A vibration sensor 22 mounted on a motor 2 is used to detect vibrations around a motor rotating shaft. A sampling pulse is generated by a rotary encoder 23 each time the motor rotating shaft 4 is rotated by a certain small angle. Time-series vibration signals outputted from the vibration sensor 22 in synchronization with the sampling pulses outputted from the rotary encoder 23 are converted from analog to digital form by an A/C convert 24 and are thereby converted into signals of rotating shaft rotary angle space. Further, a noise eliminating portion 26 subjects the rotary angle space signals to moving geometric averaging or moving arithmetical averaging, thereby eliminating noise components.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality diagnostic device for rotating equipment. In particular, the present invention converts a vibration signal obtained by measuring vibration of a rotating device into a vibration signal in an angular space about a rotation angle of a specific rotating portion of the rotating device, and analyzes the vibration signal to detect an abnormality of the rotating device. The present invention relates to an abnormality diagnosis device for recognition and diagnosis.

[0002]

2. Description of the Related Art Conventionally, in an abnormality diagnosing apparatus using vibration, a vibration signal s (t) itself which is a time series signal, and a Fourier transform signal S (f; which is obtained by converting the vibration signal s (t) into a frequency space signal. The features of t) are classified to recognize and diagnose abnormalities. FIG. 1 is a block diagram showing such a conventional abnormality diagnosing device, which is a device for diagnosing an abnormality around a shaft of a motor 2 driving a motor driving machine 1. Here, as the abnormality around the axis, for example, the motor 2
Rotor imbalance, gear 3 meshing failure, gear 3 scratches (that is, scratches, chips, wear, uneven wear, scratches in a broad sense such as defective machining), bearing inner ring scratches, bearing outer ring scratches, bearings There are scratches on rolling elements (balls, rollers).

The motor drive machine 1 is composed of a motor 2 and a gear train connected to its rotating shaft 4, and an acceleration sensor 5 is attached to the motor 2. Then, from the motor driving machine 1 side to the abnormality diagnosing device 6 side, a signal indicating the vibration of the motor 2 detected by the acceleration sensor 5, that is, a time series vibration signal s (t) is output. The abnormality diagnosis device 6 includes an amplifier 7 and a frequency filter 8
And a Fourier transform unit (FFT) 9, two feature extraction units 10 and 11, and a classification / diagnosis unit 12. Time series vibration signal s transmitted from the acceleration sensor 5
After (t) is amplified by the amplifier 7, only the signal in the frequency region required by the frequency filter 8, that is, the signal in the frequency region of the motor rotation speed and the abnormal vibration of the motor driving machine 1 is passed. Next, the signal that has passed through the frequency filter 8 is directly input to the feature extraction unit 11 on the one hand, and fast Fourier transformed to a signal in the frequency space by the Fourier transform unit 9 on the other hand, and then input to the feature extraction unit 10.

For example, the time-series vibration signal s (t) output from the acceleration sensor 5 is usually due to the imbalance of the rotor of the motor 2, and therefore, as shown in FIGS. The basic vibration signal 13 having the same frequency as the rotation speed appears. Further, when a part of the first-stage gear 3 attached to the rotary shaft 4 of the motor 2 has a scratch, FIG.
As shown in, the abnormal vibration signal 14a caused by the scratch of the gear 3 of the first stage is generated on the basic vibration signal 13 at the same cycle as the motor rotation cycle. If the gear 3 with 1/2 the number of teeth meshing with the first-stage gear 3 has a flaw, the gear 3 is damaged on the basic vibration signal 13 as shown in FIG. 2B. The abnormal vibration signal 14b
It appears in the cycle of. Similarly, when the bearing inner ring or the bearing outer ring is damaged, an abnormal vibration signal is generated at a constant cycle determined by the size or the like. Further, the time series vibration signal s (t) including the information indicating the abnormality of each of these parts is Fourier-transformed by the Fourier transform part 9 into the Fourier transform signal S in the frequency space.
When converted into (f; t), as shown in FIG.
Abnormal vibration signal 13 indicating basic vibration signal 13 and other abnormalities
4a, 14b, ... Appear at constant frequency positions in the frequency space.

However, the characteristic extracting unit 11 analyzes a large number of signals appearing in the time-series vibration signal s (t) to analyze the basic vibration signal 13 and the abnormal vibration signals 14a and 14a.
Feature extraction of b, ... Is performed. Further, the feature extraction unit 10 extracts the features of the basic vibration signal 13 and the abnormal vibration signals 14a, 14b, ... By analyzing a large number of signals represented by the Fourier transform signal S (f; t). When the feature extraction is performed, the classification / diagnosis unit 12 classifies the position of occurrence of an abnormality such as a flaw or foreign matter and its type, and outputs a diagnosis result (diagnosis output).

[0006]

The conventional abnormality diagnosis device 6
Then, when the motor 2 is rotating at a constant number of revolutions, if there is an abnormality around the axis such as an imbalance of the rotor or scratches on the gear 3, for example, as shown in FIGS. like,
The basic vibration signal 13 having a constant cycle is added to the time-series vibration signal s (t).
And abnormal vibration signals 14a, 14b, ... Appear. Also,
As shown in FIG. 2C, the Fourier transform signal S (f;
The basic vibration signal 13 and abnormal vibration signals 14a, 14b, ... Appear at the constant frequency position of t).

However, when the rotation speed of the motor 2 is changing, for example, as shown in FIG. 4A, when the motor rotation speed is increasing (during acceleration), the time series vibration signal s ( In t), the cycle of the basic vibration signal 13 fluctuates, and the abnormal vibration signal 14 does not appear in a constant cycle even if it is an abnormal vibration signal due to the same cause. Therefore, the abnormality diagnosis device 6 detects an abnormality around the axis. It was difficult to do. Similarly, when the rotation speed of the motor is changing, the cycles of the basic vibration signal 13 and the abnormal vibration signal 14 are fluctuating, and thus the basic vibration signal 13 and the abnormal vibration are included in the Fourier transform signal S (f; t). The position of the vibration signal 14 changed with time, and it was difficult to detect an abnormality around the axis.

Further, even when the motor 2 is rotating at a constant speed, it takes a long time to convert the signal in the Fourier transform section 9, and the signal processing such as the Fourier transform is very complicated.

Further, even if an abnormality is detected while the rotation speed of the motor 2 is stable at a constant value, as shown in FIG. 5A, the time-series vibration signal s varies depending on the difference in the motor rotation speed.
Basic vibration signals 13a, 13b, 13c, 13 in (t)
The period of d is significantly different. Similarly, the cycle of the abnormal vibration signal due to the damage of the gear 3 and the like greatly changes depending on the motor rotation speed. Furthermore, since the cycle of the time-series vibration signal s (t) varies, the positions at which the fundamental vibration signal 13 and the abnormal vibration signal 14 occur in the Fourier transform signal S (f; t) are also undefined depending on the motor rotation speed. . For this reason, in the motor drive machine 1 that rotates at an arbitrary number of revolutions, it is difficult to perform the feature extraction processing in the feature extraction units 10 and 11, and it is difficult to provide the abnormality diagnosis device 6 with versatility.

Further, since the time-series vibration signal s (t) contains vibrations at various points in the motor drive machine 1, it becomes a signal having a very large noise component when viewed only from the necessary signal. ing. As a measure for improving the S / N ratio, it is effective to use the frequency filter 8.
However, since the pass band required for the frequency filter 8 varies depending on the number of rotations of the motor 2, the effect could not be expected in reality.

The present invention has been made in view of the above-mentioned drawbacks of the conventional examples, and an object of the present invention is to convert a time-series vibration signal into a rotation angle spatial signal of a rotary shaft for processing. To solve these problems.

[0012]

An abnormality diagnosing device for rotating equipment according to the present invention comprises a vibration sensor for detecting vibration around a rotating shaft, and a sampling signal generating means for converting a rotation angle of the rotating shaft into a sampling signal. Conversion means for sampling the signal of the vibration sensor for each sampling signal of the sampling signal generating means and converting it into a signal of the rotation axis rotation angle space, and based on the signal of the rotation axis rotation angle space output by the conversion means. It is characterized by diagnosing abnormalities of rotating equipment.

According to the abnormality diagnosing device of the present invention, it is possible to diagnose an abnormality such as an imbalance of the center of gravity around the rotating shaft, a gear meshing failure, and a bearing scratch.

As the sampling signal generating means, it is possible to use a rotary encoder which detects a rotation angle and a rotation direction of a rotating shaft or a rotating portion which rotates in conjunction with the rotating shaft. Alternatively, sampling signal generating means may be used which detects the teeth of the gear that rotates in conjunction with the rotating shaft with a detector such as a proximity switch or an electromagnetic pickup, and thereby generates a sampling signal.

Further, the sampling signal generating means may detect the zero-cross point of the unbalanced signal due to the rotor in the signal of the vibration sensor and generate the sampling signal accordingly. Alternatively, the sampling signal generating means may be one that generates a sampling signal by the gear meshing signal in the signal of the vibration sensor.

As the timing of the diagnostic processing, the diagnostic processing may be performed at regular intervals. Alternatively, the diagnosis process may be performed when the rotation speed of the rotary shaft is a constant rotation speed. Alternatively, the diagnostic process may be performed when the timing signal is received from the programmable logic controller or other device.

[0017]

In the abnormality diagnosing device of the present invention, the vibration sensor detects the vibration around the rotary shaft, and the sampling signal generating means generates the sampling signal according to the rotation angle of the rotary shaft. The output time series vibration signal is sampled in synchronization with the sampling signal,
Thereby, the signal of the vibration sensor is converted into the signal of the rotation axis rotation angle space. Then, the abnormality diagnosis device diagnoses the abnormality of the rotating device based on the signal of the rotation axis rotation angle space.

In the rotating shaft rotation angle space, a signal due to an imbalance around the rotating shaft and a signal due to a flaw or the like around a rotating shaft have a constant cycle regardless of the rotating speed and the rotating acceleration of the rotating shaft. Becomes Further, even if the number of rotations of the rotary shaft is different, the signals have the same cycle. Therefore,
It is possible to simplify signal processing and determination processing for diagnosing an abnormality around the rotation axis.

If a rotary encoder is used as the sampling signal generating means, the sampling signal can be easily generated only by providing the rotary encoder on the rotary shaft or the like.

Further, if the teeth of the gear that rotates in conjunction with the rotary shaft are detected by a detector such as a proximity switch or an electromagnetic pickup, the sampling signal is multiplied by multiplying the output of the detector as necessary. Can be generated. Moreover, since it is not necessary to incorporate the proximity switch and the electromagnetic pickup into the mechanism around the rotary shaft, the sampling signal generating means can be retrofitted.

Further, as the sampling signal generating means, a sampling signal is generated by detecting the zero-cross point of the unbalanced signal by the rotor in the signal of the vibration sensor or by using the gear meshing signal in the signal of the vibration sensor. By doing so, the sampling signal can be generated electrically.

[0022]

FIG. 3 is a block diagram showing an abnormality diagnosing apparatus 21 according to an embodiment of the present invention, which is an apparatus for diagnosing an abnormality around a shaft of a motor 2 which drives a motor driving machine 1. . A vibration sensor 22 is attached to the motor drive machine 1 at a position where it is easy to detect vibration around the rotation axis of the motor 2, preferably inside or outside of the motor 2 itself. The time series vibration signal s (t) generated by is output. Here, as the vibration sensor 22, various sensors such as an acceleration sensor, a speed sensor, or a position sensor that detects a minute displacement can be used. The number of vibration sensors 22 attached to the motor drive machine 1 or the motor 2 may be one, but a plurality of vibration sensors 22 of the same type or different types may be attached. When a plurality of vibration sensors 22 are attached, they may be arranged so that they have sensitivities orthogonal to each other. In particular, it is preferable to attach three vibration sensors 22 of the same type or different types to the motor 2 so that the respective sensitivity directions are orthogonal to each other.

Further, the motor rotating shaft 4 or the motor rotating shaft 4
A rotary encoder 23 is provided as a pulse signal generating means on another rotary shaft that rotates in conjunction with. The rotary encoder 23 detects the rotation angle θ of the motor rotation shaft 4 and outputs a sampling pulse P every time the motor rotation shaft 4 rotates by a fixed minute angle Δθ. As the rotary encoder 23, it is preferable to use an incremental rotary encoder capable of detecting the rotation angle θ of the motor rotating shaft 4 and also detecting the rotation direction of the motor rotating shaft 4, and the mounting position is also provided directly on the motor rotating shaft 4. Is preferred. If the incremental rotary encoder is used, the direction of rotation of the motor rotary shaft 4 can be determined, and the sampling pulse P is output not only when the rotary shaft 4 is rotating normally but also when the rotary shaft 4 is rotating in the reverse direction. Can be output. Furthermore, if an incremental rotary encoder that can output a zero-point signal is used, the origin of the rotation axis rotation angle space can always be fixed.

The abnormality diagnosing device 21, as shown in FIG.
Analog-to-digital conversion circuit (hereinafter A)
/ D converter) 24, a plurality of m-ary counters 25 that can set the count value m, a plurality of noise removal units 26 that perform moving average processing, a feature extraction unit 27, and a classification diagnosis unit 2
It is composed of 8. The vibration around the motor rotation axis detected by the vibration sensor 22 is input to the A / D converter 24 as a time series vibration signal s (t). On the other hand, in the rotary encoder 23, the motor rotation shaft 4 has a constant small angle Δ.
The sampling pulse P is generated each time θ rotation is performed.
The A / D converter 24 converts the time series vibration signal s (t), which is an analog signal, into the rotation angle space signal R (θ), which is a digital signal. That is, the A / D converter 24
Then, the time-series vibration signal s (t) is sampled in synchronization with the sampling pulse P output from the rotary encoder 23, and is output as a multilevel digital signal (for example, 256-level 8-bit signal). Sampling pulse P
Is output every time the motor rotating shaft 4 rotates by a constant angle, and the multi-valued digital signal becomes a signal for each constant rotating angle of the motor rotating shaft 4, and thus is a signal in the rotating shaft rotating angle space. That is, the time series vibration signal s (t) is converted into the rotation angle space signal R (θ) by the A / D converter 24.

For example, consider a case where the rotation speed of the motor 2 increases with time. In this case, in the time series vibration signal s (t) shown in FIG. 4A, the basic vibration signal 13 and the gear 3 generated by the imbalance of the rotor of the motor 2 are generated.
The abnormal vibration signal 14, which is generated due to scratches or the like, is generated at an irregular cycle. However, at this time, since the sampling pulse P is generated from the rotary encoder 23 at a frequency proportional to the motor rotation number as shown in FIG. 4B, the time series vibration signal s (t) is generated using this sampling pulse P. Is the rotation angle space signal R in the rotation axis rotation space.
When converted into (θ), the basic vibration signal 13 and the abnormal vibration signal 14 appear in a constant cycle as shown in FIG. 4 (c). Therefore, if feature extraction or the like is performed using this rotation angle spatial signal R (θ), feature extraction or the like is performed using the time-series vibration signal s (t) or its Fourier transform signal S (f; t) as in the conventional case. The signal processing can be extremely facilitated as compared with the case where

Further, consider a case where the rotation speed of the motor 2 is constant, but the rotation speed of the motor 2 is different for each operation. FIG. 5A is a diagram showing the time-series vibration signal s (t), and shows each basic vibration signal 13 when the rotation speed of the motor 2 is different. The motor rotation speed increases in the order of the basic vibration signals 13a, 13b, 13c, 13d. These basic vibration signals 13a, 13b, 13
When c and 13d are converted into the rotation angle space signal R (θ) by the A / D converter 24, the original fundamental vibration signal 13
Each of a, 13b, 13c, and 13d becomes the basic vibration signal 13 having the same period as shown in FIG. 5B. In particular, the cycle of the basic vibration signal 13 in the rotation angle space signal R (θ) does not change depending on the rotation speed or the rotation acceleration of the motor 2 and is always 2π radians. Therefore, also in this case, if the feature extraction or the like is performed using the rotation angle spatial signal R (θ), the feature is obtained using the time-series vibration signal s (t) and its Fourier transform signal S (f; t) as in the conventional case. Compared with the case of performing extraction or the like, signal processing becomes simpler.

When the time series vibration signal s (t) is converted into the rotation angle space signal R (θ) in this way, the vibration signal due to the same cause is irrespective of the rotation speed and the rotation acceleration of the motor 2. , In the rotation axis rotation angle space, it always appears with a constant period. For example, the basic vibration signal 13 due to the imbalance of the rotor of the motor 2 always appears in a cycle of 2π radians as described above. Further, as shown in FIG. 6, the vibration (abnormal vibration signal) due to the meshing of the first-stage gear 3 attached to the motor rotation shaft 4 has a period of 2π on the basic vibration signal 13 due to the imbalance of the rotor.
/ K radian (K is the number of teeth of the first-stage gear 3) is represented as an abnormal vibration signal 14A.

Further, when the inner ring of the rolling bearing (not shown) that rotatably supports the motor rotation shaft has one flaw, the diameter of the inner ring raceway is Di and the outer ring raceway diameter is Do. When the number of rolling elements (balls, rollers; assumed to be evenly arranged by separators) is N, the cycle Ti of the abnormal vibration signal 14B when the scratches on the inner ring contact the rolling elements
Is Ti = 2π (Di + Do) / (DoN) radians. Now, assuming that the diameter of the inner ring raceway Di = 50 mm, the diameter of the outer ring raceway Do = 80 mm, and the number of rolling elements N = 6, the cycle Ti = 0.54π of the abnormal vibration signal 14B when the inner ring scratch contacts the rolling elements. Become radians. Therefore, as shown in FIG. 7, the abnormal vibration signal 14B appears on the basic vibration signal 13 at a constant cycle determined by the dimensions of the rolling bearing.

When the outer ring of the rolling bearing has one flaw, the cycle To of the abnormal vibration signal 14C when the flaw of the outer ring contacts the rolling element is To = 2π (Di + D).
o) / (DiN) radians. Also using the above numerical values, the cycle To of the abnormal vibration signal 14C when the scratch on the outer ring contacts the rolling element is To = 0.86π radian. Therefore, also in this case, as shown in FIG.
The abnormal vibration signal 14C appears on the upper surface of the rolling bearing at a constant period determined by the dimensions of the rolling bearing.

When the rolling element of the rolling bearing has one flaw, the cycle Tir of the abnormal vibration signal 14D when the flaw of the rolling element comes into contact with the inner ring is given by the diameter d of the rolling element. , Tir = 2πd (Di + Do) / (DiDo
N) Radians. Similarly, the cycle Tor of the abnormal vibration signal 14D when the wound of the rolling element contacts the outer ring is Tor.
= 2πd (Di + Do) / (DiDoN) radians. Also using the above numerical values, the cycle of the abnormal vibration signal 14D when the scratch of the rolling element contacts the inner ring or the outer ring is
Tir = Tor = 0.98π radian (phase difference 0.6π radian). Therefore, also in this case, as shown in FIG. 9, the abnormal vibration signal 14D appears on the basic vibration signal 13 at a constant cycle determined by the dimensions of the rolling bearing.

A follower gear separated from the motor rotation shaft,
Although load fluctuations, noise, etc. of the driven shaft having the driven gear occur asynchronously with the rotation angle of the motor rotation shaft, this abnormality diagnosis device uses a vibration sensor for the purpose of abnormality diagnosis around the motor rotation shaft. Since the motor rotation shaft is provided at a position where vibration can be easily detected, these asynchronous signals are treated as noise that is not a diagnosis target. Also, these are relatively small signal levels.

Next, noise removal by moving average using the m-ary counter 25 and the noise removal unit 26 will be described. The m-ary counter 25 starts counting from 0, outputs a carrier signal when the count value reaches m, and clears the carrier signal to 0. The count value m can be arbitrarily set in advance. Each m-ary counter 25 counts in synchronization with the sampling pulse output from the rotary encoder 23. Therefore, when the number of sampling pulses output from the rotary encoder 23 during one rotation of the motor rotation shaft 4 is M, the m-ary counter 25 outputs a carrier signal at a cycle of 2π (m / M) radian. That is, if the count value m of each m-ary counter 25 is set to n1, n2, ..., From each m-ary counter 25, 2π (n1 / M) radian,
A carrier signal is output in each cycle of 2π (n2 / M) radians, ....

The plurality of noise removing units 26 are respectively A /
The rotation angle space signal R converted by the D converter 24
(Θ) is input, and each noise removing unit 26 performs a moving average of the rotation angle spatial signals R (θ) at different sampling periods (for example, moving geometric average, moving arithmetic average, etc.).
By doing so, noise or a signal having a cycle different from each sampling cycle is removed as a noise component. That is, each m-ary counter 2 corresponding to each noise removing unit 26
5 carrier signals are input, and each noise removal unit 2
6 performs a moving average process in synchronization with the inputted carrier signal. For example, in the noise removing unit 26 connected to the m-ary counter 25 having the count value m = nj, 2π (nj
/ M) A moving geometric average or a moving arithmetic average is performed at a sampling period of radian. Therefore, it is possible to output the rotation angle spatial signal R (θ) having a high S / N ratio from which the noise component is removed from each noise removal unit 26 to the feature extraction unit 27, and the signal component can be made clear. . In addition,
The moving geometric mean or the moving arithmetic mean in each noise removing unit 26 is executed with a phase shift.

For example, when diagnosing a flaw on the inner ring of the rolling bearing, the period Ti of the abnormal vibration signal 14B generated by the flaw on the inner ring is calculated or experimentally obtained, and one m-ary counter 25 Count value m = m
Set so that MTi / (2π). As shown in FIG. 10, the corresponding noise removing unit 26 rotates the rotation angle spatial signal R with a period 2πm / M = Ti equal to the period (Ti = 0.54π) of the abnormal vibration signal 14B due to the inner ring flaw.
(Θ) is sampled, and this is subjected to moving geometric mean or moving arithmetic mean. Abnormal vibration signal 14B due to inner ring damage
Is equal to the sampling period of the noise removing unit 26, and therefore remains as it is even if the moving geometric average or the mobile arithmetic average is performed. However, noise or an abnormal vibration signal having a different period is subjected to the moving geometric average or the mobile arithmetic average, and Treated as an ingredient and suppressed. Therefore, the noise removing unit 26 extracts only the abnormal vibration signal 14B due to the scratch on the inner ring.

Similarly, in the other m-ary counter 25 and the noise removing section 26, the cycle To of the abnormal vibration signal 14C generated by the scratch on the outer ring is calculated or experimentally obtained, and the count of the m-ary counter 25 is counted. The value m is m = M
Set so that it becomes To / (2π). As shown in FIG. 11, the corresponding noise removing unit 26 rotates the rotation angle spatial signal R (θ) at a period 2πm / M = To equal to the period (To = 0.86π) of the abnormal vibration signal 14C due to the outer ring flaw.
Is sampled and this is subjected to a moving geometric mean or a moving arithmetic mean. Thereby, from the noise removing unit 26,
The noise is suppressed and the abnormal vibration signal 14C due to the damage to the inner ring is output at a high S / N ratio.

Further, in another m-ary counter 25 and noise removing section 26, the period Tir = Tor of the abnormal vibration signal 14D generated by the scratch of the rolling element is calculated or experimentally obtained, and the m-ary counter 25 The count value m is set so that m = MTir / (2π) = MTor / (2π). The corresponding noise removing unit 26, as shown in FIG. 12, receives the abnormal vibration signal 14D due to the scratch on the rolling element.
Period equal to the period (To = 0.98π) 2πm / M =
The rotation angle spatial signal R (θ) is sampled at Tir = Tor, and this is subjected to a moving geometric mean or a moving arithmetic mean. As a result, the noise removal unit 26 suppresses noise and outputs the abnormal vibration signal 14D due to the scratch on the rolling element at a high S / N ratio.

Similarly, by setting the count value m corresponding to the cycle in the rotation angle space of the abnormal phenomenon to be detected in each of the other m-ary counters 25 and the noise removing section 26, Abnormalities can be detected.

In the feature extraction unit 27, each noise removal unit 26
Feature extraction is performed on the signal that has been subjected to the moving average processing. In the feature extraction unit 27, values such as threshold values, statistical quantities (average value, variance, maximum value, minimum value, correlation coefficient), discriminant function, and the like, as in the method used in the image analysis method, Feature extraction is performed using the quantity. When the feature extraction is performed in this way, the classification / diagnosis unit 28 uses the maximum likelihood method, the shortest distance classification method, the cluster analysis method, and the like, which are also used in the image analysis method, to detect scratches, chips, foreign matter, and the like. Such an abnormality occurrence position and type are classified and a diagnosis result (diagnosis output) is output.

As described above, the abnormality diagnosis device 21 of the present invention.
According to the above, since the time series vibration signal s (t) is converted into the rotation angle space signal R (θ) to perform the abnormality diagnosis, the motor 2 changes when the motor rotation speed changes or at a different rotation speed. Even when rotating, the abnormal vibration signal can be inspected as the vibration signal of a constant cycle, and the abnormality diagnosis can be performed easily and surely. Further, in the noise removal unit 26, the rotation angle spatial signal R (θ) is calculated using the moving average.
Since signals other than the fixed period of are removed, high S / N
A signal having a ratio can be obtained, and highly accurate abnormality detection can be performed. Moreover, unlike the frequency filter of the conventional example, there is no problem that the required characteristics change even if the rotation speed of the motor 2 changes.

FIG. 10 is a block diagram showing an abnormality diagnosis device 29 according to another embodiment of the present invention. In this abnormality diagnosing device 29, a sampling signal is generated by a detector 30 for detecting the teeth of the first-stage gear 3 attached to the motor rotating shaft 4, particularly a metal sensor such as a proximity switch or an electromagnetic pickup, and a frequency multiplication circuit 31. Constitutes a means. The detection signal, which is detected by the detector 30 to detect the teeth of the gear 3 and is output, is generated every time the motor rotation shaft 4 rotates by a constant angle, and becomes a signal of a constant cycle in the rotation angle space (FIG. 9). The frequency multiplication circuit 31 converts the signal into several tens to several 1
A sampling signal can be obtained by multiplying by 00 and outputting.

FIG. 11 is a block diagram showing an abnormality diagnosing device 32 according to still another embodiment of the present invention. In the abnormality diagnosis device 32, the zero-cross point detector 33 that detects the zero-cross point of the basic vibration signal 13 included in the time-series vibration signal s (t) output from the vibration sensor 22 and the frequency multiplication circuit 31 It constitutes a sampling signal generating means. The basic vibration signal 13 included in the time-series vibration signal s (t) changes with the rotation angle of the motor rotation shaft 4 regardless of the rotation speed and the rotation acceleration of the motor rotation shaft 4, and has a constant cycle in the rotation angle space. (Fig. 4,
From FIG. 5), this detection signal is converted into
It can be used as a sampling signal by multiplying by 0 to several hundreds and outputting.

Further, since the basic vibration signal 13 of the time-series vibration signal s (t) carries the abnormal vibration signal 14A due to the meshing of the first-stage gear 3 attached to the motor rotation shaft 4, (the rotation angle spatial signal R By detecting the abnormal vibration signal 14A due to the meshing of the first stage gear 3, a sampling signal corresponding to the rotation angle of the motor rotation shaft 4 can be obtained. For example,
In the abnormality diagnosis device 34 shown in FIG. 12, the basic vibration signal 13
Of the frequency band of the abnormal vibration signal 14A caused by the meshing of the first-stage gear 3 and a zero-pass point detector 33 for detecting the zero-cross point of the signal after passing through the high-pass filter 35.
And the frequency multiplication circuit 31 constitute a sampling signal generating means. When the basic vibration signal 13 of the time-series vibration signal s (t) is cut by the high-pass filter 35, the time-series vibration signal s (t) after passing through the high-pass filter 35 is zero-crossed by the abnormal vibration signal 14A due to the meshing of the first gear 3. It will be. Therefore,
By detecting the zero-cross point of the signal by the zero-cross point detector 33 and multiplying it by several tens to several hundreds by the frequency multiplication circuit 31, the target sampling signal can be obtained as in the embodiment of FIG.

The abnormality diagnosing device may always be made to perform the abnormality diagnosis, but the abnormality diagnosis may be made only when necessary or suitable for the abnormality diagnosis. For example, the abnormality diagnosis device may perform the abnormality diagnosis only when the motor is rotating at a constant rotation speed. In the embodiment shown in FIG. 13, the rotational speed of the motor rotating shaft 4 is monitored by the sampling pulse output from the rotary encoder 23 in the rotating speed detecting unit 37, and the rotating speed of the motor rotating shaft 4 is at a constant speed. The abnormality diagnosing device 36 is turned on at the time of, and the abnormality diagnosing device 36 is turned off when the rotation speed fluctuates or becomes unstable. If the abnormality diagnosis is performed only when the motor rotation speed is stable in this way, the accuracy of the diagnosis result can be increased.

Further, the abnormality diagnosis may be performed for a fixed time. For example, as in the embodiment shown in FIG. 14, the timer 38 is set to the on-time and the off-time of the abnormality diagnosing device 36 so that the abnormality diagnosing can be repeatedly performed at regular intervals. As a result, it is possible to periodically diagnose the abnormality instead of continuously performing the abnormality diagnosis. Alternatively, when the operator turns on the timer 38, the abnormality diagnosis device 36 is turned on, and when the set time of the timer 38 has elapsed, the abnormality diagnosis device 36 is turned off again. By doing so, it is possible to appropriately perform the abnormality diagnosis around the motor shaft only when necessary.

Further, as shown in FIG. 15, programmable logic controller (PCL) and other equipment 39,
For example, the diagnosis process may be performed by outputting a timing signal from the control device that controls the motor drive machine 1. In this case, for example, when the motor drive machine 1 is started, the abnormality diagnosis of the motor drive machine 1 can be automatically performed.

In each of the above embodiments, the abnormality diagnosis device is used for the motor drive machine, but the drive source is not limited to the motor (electric motor), and if the reciprocating engine is focused on the rotating shaft, it is driven. The abnormality diagnosis device of the present invention can be applied to any rotation shaft from the shaft to the output shaft.

[0047]

According to the present invention, the time-series vibration signal output from the vibration sensor is converted into a signal of the rotating shaft rotation angle space, and the abnormality of the rotating device is diagnosed based on this rotation angle space signal. Therefore, the abnormality diagnosis can be performed by the rotation angle spatial signal having a constant cycle irrespective of the rotation speed and the rotation acceleration of the rotation shaft. Further, even if the number of rotations of the rotation shaft is different, the rotation angle space signal has the same period. Therefore, it is possible to classify the types of abnormality around the rotation axis into patterns according to the period of the signal in the rotation axis rotation angle space, and to simplify the signal processing and the judgment processing for failure diagnosis.

Further, since the Fourier transform is not required unlike the conventional example, the signal processing time can be shortened and the signal processing can be simplified.

Further, since the rotation angle spatial signal becomes a signal having a constant cycle, it is possible to use noise averaging means such as moving geometric mean or moving arithmetic mean for noise elimination. Unlike the conventional frequency filter, this noise removing means does not need to change the frequency characteristic according to the rotation speed, the rotation acceleration, or the like, can achieve a high S / N ratio, and can improve the accuracy of abnormality detection. .

If a rotary encoder is used as the sampling signal generating means, the sampling signal can be easily generated only by providing the rotary encoder on the rotary shaft or the like. Moreover, if an incremental rotary encoder that can detect the direction of rotation is used, abnormality diagnosis processing can be performed regardless of whether the rotation axis is rotating in the forward or reverse direction. Further, if the teeth of the gear are detected by a detector such as a proximity switch or an electromagnetic pickup, the sampling signal can be generated by multiplying the output of the detector as needed.
Moreover, when a proximity switch or an electromagnetic pickup is used, it is not necessary to incorporate them into a mechanism around the rotation axis, so that the sampling signal generating means can be retrofitted.
Further, as the sampling signal generating means, the sampling signal is generated by detecting the zero crossing point of the unbalanced signal by the rotor in the signal of the vibration sensor or by using the gear meshing signal in the signal of the vibration sensor. Then, the sampling signal can be electrically generated.

Further, if the abnormality diagnosis processing is performed at regular intervals, the abnormality diagnosis can be performed periodically. Alternatively, if the abnormality diagnosis processing is performed when the rotation speed of the rotating shaft is constant, the abnormality diagnosis processing can be performed in a stable state, the diagnosis processing can be facilitated, and the accuracy of the diagnosis result is improved. To do. Further, if a timing signal is obtained from a programmable logic controller or other device to perform abnormality diagnosis processing, it is possible to perform abnormality diagnosis, for example, at the start of operation of a rotating device,
The rotating equipment can be operated after confirming that there is no abnormality.

[Brief description of drawings]

FIG. 1 is a block diagram showing a conventional abnormality diagnosis device.

2A and 2B are waveform diagrams showing a time-series vibration signal of a rotary shaft, and FIG. 2C is a waveform diagram showing a Fourier transform signal converted from the time-series vibration signal.

FIG. 3 is a block diagram showing an abnormality diagnosis device in one embodiment of the present invention.

4A is a waveform diagram showing a time-series vibration signal when the motor rotation speed is increasing, FIG. 4B is a diagram showing sampling pulses in the same state, and FIG. 4C is a converted rotation. It is a wave form diagram which shows an angle space signal.

5A is a waveform diagram showing a time-series vibration signal in each state where the motor rotation speed is different, and FIG. 5B is a waveform diagram showing a rotation angle space signal converted from each time-series vibration signal.

FIG. 6 is a waveform diagram showing another example of a rotation angle space signal.

FIG. 7 is a waveform chart showing still another example of the rotation angle spatial signal.

FIG. 8 is a waveform diagram showing still another example of the rotation angle spatial signal.

FIG. 9 is a waveform diagram showing still another example of the rotation angle spatial signal.

FIG. 10 is a block diagram showing an abnormality diagnosis device according to another embodiment of the present invention.

FIG. 11 is a block diagram showing an abnormality diagnosis device according to still another embodiment of the present invention.

FIG. 12 is a block diagram showing an abnormality diagnosis device in still another embodiment of the present invention.

FIG. 13 is a schematic block diagram showing an abnormality diagnosis device according to still another embodiment of the present invention.

FIG. 14 is a schematic block diagram showing an abnormality diagnosis device in still another embodiment of the present invention.

FIG. 15 is a schematic block diagram showing an abnormality diagnosis device according to still another embodiment of the present invention.

[Explanation of symbols]

 2 Motor 3 Gear 4 Rotating shaft of motor 13 Vibration signal due to unbalance of rotor 14, 14A to 14D Abnormal vibration signal 22 Vibration sensor 23 Rotary encoder 24 A / D converter 25 M-ary counter 26 Noise eliminator 30 Proximity switch and electromagnetic pickup Detectors 31 Frequency multiplication circuit 33 Zero cross point detector

Claims (9)

[Claims] 1. A vibration sensor for detecting a vibration around a rotation axis, a sampling signal generating means for converting a rotation angle of the rotation axis into a sampling signal, and a signal of the vibration sensor for each sampling signal of the sampling signal generating means. An abnormality diagnosing device for a rotating device, comprising: a converter for sampling and converting the signal into a signal of the rotating shaft rotating angle space; and diagnosing an abnormality of the rotating device based on the signal of the rotating shaft rotation angle space output from the converter. 2. The abnormality diagnosing device for a rotating machine according to claim 1, wherein an abnormality such as an imbalance of a center of gravity around the rotating shaft, a gear meshing failure, and a bearing scratch is diagnosed. 3. The rotary encoder according to claim 1, wherein the sampling signal generating means is a rotary encoder that detects a rotation angle and a rotation direction of a rotating shaft or a rotating portion that rotates in conjunction with the rotating shaft. Abnormality diagnosing device for rotating equipment. 4. The sampling signal generating means detects a tooth of a gear that rotates in conjunction with a rotating shaft with a detector such as a proximity switch or an electromagnetic pickup, and thereby generates a sampling signal. The abnormality diagnosis device for a rotating device according to claim 1 or 2. 5. The sampling signal generating means detects a zero cross point of an unbalanced signal due to a rotor in a signal of a vibration sensor, and generates a sampling signal by the zero crossing point. An apparatus for diagnosing abnormalities in rotating equipment according to item 1. 6. The sampling signal generating means generates a sampling signal according to a gear meshing signal in the signal of the vibration sensor.
Alternatively, the abnormality diagnosis device for rotating equipment according to item 2. 7. The abnormality diagnosing device for a rotating machine according to claim 1, 2, 3, 4, 5 or 6, wherein the diagnosing process is performed at regular intervals. 8. The rotation according to claim 1, 2, 3, 4, 5 or 6, wherein the diagnosis processing is performed when the rotation speed of the rotary shaft is a constant rotation speed. Equipment abnormality diagnosis device. 9. The diagnostic processing is performed by obtaining a timing signal from a programmable logic controller or other device.
4. An abnormality diagnostic device for rotating equipment according to 4, 5, or 6.