TWI414919B - Early-warning apparatus for health detection of servo motor and method for operating the same - Google Patents

Abstract

An early-warning apparatus for health detection of a servo motor and a method for operating the same are applied to estimate vibration phenomenon of a CNC tool machine. First, a vibration signal is produced through a vibration detecting unit. Afterward, the vibration signal is sequentially sent to a data buffer. Afterward, the vibration signal is transformed in time and frequency domains through a time-frequency transforming unit. Finally, a deterioration index is calculated through a deterioration index calculating unit to obtain a health index which is calculated through a health index calculating unit. Therefore, the build-in vibration detecting unit is provided to dispense with additional external sensors. Furthermore, the vibration phenomenon of the servo motor of the CNC tool machine is estimated according to the health index, thus analyzing non-linear and non-stationary characteristics of the estimated vibration phenomenon.

Description

Servo motor health warning device and calculation method thereof

The invention relates to a health warning device for a servo motor and a calculation method thereof, and provides a health warning device for a servo motor for vibration signal time-frequency domain conversion and a calculation method thereof.

In recent years, with the improvement of high-efficiency and high-quality production technology, the mechanical equipment has been comprehensively oriented, large-scale, rapid, systematic, complicated, and automated. Due to the increasing scale of equipment in the factory, the connection between the systems is becoming more and more close, and the equipment itself is becoming more and more complicated. Therefore, if it is impossible to predict in advance the possible failures, the equipment will be repaired and maintained. Once the fault occurs, the economic losses caused are also very considerable.

Taking the operation of the computer numerical control machine tool as an example, the host computer transmits the position command to the multi-axis servo driver for driving the servo motor to rotate. Through the action of the transmission system (such as guide bolts, slide rails, etc.), the platform can be moved. However, long-term operation, mechanical wear, lubrication conditions, and the position of the heart will affect the smoothness of the machine operation. Therefore, the machine generates irregular vibration and energy loss, which cannot be avoided. In the process of equipment operation, if abnormal vibration exists in the equipment for a long time, it will not be improved immediately. Although it can still maintain normal operation in a short time, it will cause damage to the equipment and affect its working efficiency.

For systematic calculations to estimate the overall servo motor health index, a PC-based calculator is used to collect the voltage, current, lossy losses of the driver and the vibration values transmitted from the accelerometer on the machine. These time domain data can be replaced by a Fast Fourier Transform (FFT) or a Wavelet Transform (WT) into a frequency domain or a time-frequency domain. However, these time-frequency domain conversion techniques, although using statistical mathematics and pattern learning, can obtain current health status indicator values. However, because of the large amount of calculation, a separate calculator is required to perform, and thus, the cost and space for installing the calculators will be increased. At the same time, due to the differences in the driver functions of different brands, the source of the signals that can be captured and the immediacy will be limited.

Research on fault diagnosis of rotating machinery has been developed for many years. So far, there are many methods for detecting vibration signals. For signal processing and data analysis, Fast Fourier Transform (FFT) is the most commonly used method at present, and it is the frequency domain. The most representative one of the analytical methods. Traditional Fourier spectral analysis provides a convenient way to analyze the energy distribution of a data in a frequency domain based on the composition of the signal being linearized by sine or cosine functions of different frequencies, amplitudes, and phases. The combination of superposition makes signal features that are difficult to express in the time domain clearly visible in the frequency domain. As long as the signal is a stationary distribution and a linear time series, the characteristics of the signal can be effectively transmitted through the spectral conversion. However, for non-linear and nonstationary data, Fourier analysis mainly causes the following shortcomings:

1. It is easy to erase some messages during the integration process, and it is also possible to generate the spectrum illusion by spreading the energy to the high frequency part due to the integration, so that the correct frequency Fuzzification, causing errors in interpretation.

2. When the signal is converted into the frequency domain, the information in the time domain disappears, that is, the time at which the specific frequency occurs cannot be determined in the frequency domain, and the inconvenience in analyzing the signal is increased.

For wavelet transform (WT), the distribution of the signal in the three-dimensional space of energy-frequency-time can be analyzed. It can decompose the mixed signal composed of different frequencies into signals of different frequency components, which can effectively separate the signal and noise. However, since the wavelet transform is corrected by the Fourier transform, there are still disadvantages of energy dispersion, increased bandwidth, and lack of adaptability. Also, the appropriate wavelet basis function must be selected before the signal is decomposed. Once this basis function is selected, it must be used to analyze all the data, thus limiting the applicable range.

Therefore, how to design a health warning device for servo motor and its calculation method can reduce the installation and wiring of the external sensor and provide better analysis of the nonlinear and non-steady-state vibration characteristics. A major issue that people want to overcome and solve.

In order to solve the above problems, the present invention provides a health warning device for a servo motor, which is applied to the vibration condition estimation of a computer numerical control tool machine.

The servo motor's health warning device includes a servo motor and a servo drive. The servo motor has at least one vibration sensing unit built therein to sense the operating parameters of the servo motor and generate a vibration signal.

The servo drive is connected to the servo motor and contains a microprocessor. Microprocessor system contains Time-frequency domain conversion unit, analysis unit, deterioration index calculation unit, and health index calculation unit. The time-frequency domain conversion unit receives the vibration signal and provides a conversion of the vibration signal between the time domain and the frequency domain. The parsing unit is connected to the time-frequency domain converting unit to receive the vibration signal and decompose the vibration signal into a plurality of decomposition signals. The deterioration index calculation unit is connected to the analysis unit, and compares the selected decomposition signal with the evaluation curve to calculate a deterioration index. The health index calculation unit is a connection deterioration index calculation unit that calculates a health index based on the deterioration index.

By using the built-in vibration sensing unit, the installation and wiring of the additional sensor can be omitted, and the vibration condition of the servo motor running on the computer numerical control tool machine can be estimated according to the size of the health index. And provide better resolution for the nonlinear and non-steady state vibration characteristics.

In order to solve the above problems, the present invention provides a health warning calculation method for a servo motor, which is applied to the vibration condition estimation of a computer numerical control tool machine.

The steps of the health warning calculation method of the servo motor include: firstly, generating a vibration signal through the vibration sensing unit; then, sequentially transmitting the vibration signal to the data buffer; and then performing a time-frequency domain on the vibration signal through the time-frequency domain conversion unit Conversion; Finally, the deterioration index is calculated by the deterioration index calculation unit, and the health index is obtained by using the health index calculation unit.

In order to further understand the technology, the means and the effect of the present invention in order to achieve the intended purpose, refer to the following detailed description of the invention and the accompanying drawings. The detailed description is to be understood as illustrative and not restrictive.

〔this invention〕

10‧‧‧Servo motor

102‧‧‧Vibration sensing unit

20‧‧‧Servo drive

202‧‧‧High-speed serial communication interface

204‧‧‧Data buffer

206‧‧‧Microprocessor

2062‧‧‧time frequency domain conversion unit

2064‧‧‧ analytical unit

2066‧‧‧ Deterioration Index Calculation Unit

2068‧‧‧Health Index Calculation Unit

S100~S400‧‧‧Steps

S310~S340‧‧‧Steps

Sv‧‧‧ time domain vibration signal

St1~St9‧‧‧ time domain decomposition signal

△g1~△g9‧‧‧Inferior intensity

FIG. 1 is a schematic view showing a servo motor connected to a servo driver according to the present invention; FIG. 1B is a perspective view of the servo motor applied to a computer numerical control machine tool according to the present invention; The block diagram of the servo driver; the third diagram is a flowchart of the health warning calculation method of a servo motor of the present invention; the fourth diagram is a flowchart of the time-frequency domain conversion step of the health warning calculation method of the present invention; A waveform diagram of an original time domain vibration signal and a plurality of time domain decomposition signals; a fifth diagram B is an amplitude-frequency-time three-dimensional distribution map of the time domain decomposition signals of the present invention; and a sixth diagram of the present invention A schematic diagram comparing a time domain decomposition signal with a evaluation curve.

The technical content and detailed description of the present invention are as follows: Referring to FIG. 1A and FIG. B, respectively, a schematic diagram of a servo motor and a servo driver connected to the present invention and the servo motor are applied. A perspective view of a computer numerical control tool machine. Taking the machine application of the computer numerical control tool machine as an example, in the case where angular positioning is required, the motor needs an encoder to detect the rotor angle, and then estimate the angular velocity and the angular acceleration value. In actual operation, it is sometimes driven directly by a motor rotor, sometimes by a coupling. The coupling is more elastic in the direction of the rotor axis (Z direction) to deal with the movement or vibration before and after, and the X-Y plane vibration perpendicular to the axis direction Most of it will pass from the point of occurrence of the organization to the encoder. Therefore, a vibration sensing unit 102 is disposed in the encoder of the servo motor 10, wherein the vibration sensing unit 102 can be a G-sensor. The vibration sensing unit 102 detects the vibration and noise of the servo motor 10, the screw, the slide rail, and the work of the table. In addition, another vibration sensing unit 102 is fixed in the stator slot of the servo motor 10. Since the servo motor 10 is screwed onto the mechanism, the vibration of all the platforms can be accurately detected by the vibration sensing unit. 102 detected. That is to say, the vibration sensing unit 102 fixed on the encoder is used to detect the vibration wave of the transmission system; and the vibration sensing unit 102 fixed in the stator slot is used to detect the shock wave on the upper and lower platforms.

Please refer to the second figure, which is a block diagram of the servo motor and the servo driver of the present invention. The servo drive system of the computer numerical control machine tool mainly comprises a servo motor 10 and a servo driver 20. The servo motor 10 mainly includes a rotor (not shown), a stator (not shown), an encoder (not shown) mounted on the rotor, and at least one vibration sensing unit 102. The vibration sensing unit 102 can be installed in the encoder of the servo motor 10 to estimate the vibration state of the transmission system of the computer numerical control tool machine; or can be installed in the servo motor 10 In the stator slot, the computer numerical value is used to control the vibration state of the machine tool.

In actual operation, the servo motor 10 can simultaneously build a plurality of the vibration sensing units 102 in the encoder and the stator slot for respectively detecting the transmission system and the machine table of the machine tool in the X, Y, and Z directions. The vibration situation. However, for convenience of explanation, in the present embodiment, a vibration sensing unit 102 is exemplified. The vibration sensing unit 102 senses an operating parameter of the servo motor 10 and generates a vibration signal Sv. The servo driver 20 is connected to the servo motor 10. And, the servo driver 20 The system includes a high speed serial communication interface 202, a data buffer 204, and a microprocessor 206. The data buffer 204 is connected to the high-speed serial communication interface 202 for receiving and storing the vibration signal Sv generated by the vibration sensing unit 102. The data buffer 204 is a buffer (queue). Buffer).

The microprocessor 206 is coupled to the data buffer 204. The microprocessor 206 includes a time-frequency domain conversion unit 2062, a parsing unit 2064, a degradation index calculation unit 2066, and a health index calculation unit 2068. The time-frequency domain conversion unit 2062 receives the vibration signal Sv output by the data buffer 204 and provides a conversion between the time domain and the frequency domain. The parsing unit 2064 is connected to the time-frequency domain converting unit 2062 to receive the vibration signal Sv and decompose the vibration signal Sv into a plurality of decomposition signals St1 to St9 (see FIG. 5A). The deterioration index calculation unit 2066 is connected to the analysis unit 2064, and compares the selected decomposition signals St1 to St9 with a evaluation curve to calculate a deterioration index. Among them, the evaluation curve can be obtained by the empirical rule of practical operation. The health index calculation unit 2068 is connected to the deterioration index calculation unit 2066, and calculates a health index based on the deterioration index. As for the calculation of the deterioration index and the health index, it is described in detail later. Thereby, according to the size of the health index, the vibration condition of the servo motor 10 running on the computer numerical control machine tool is estimated.

Please refer to the third figure, which is a flow chart of a health warning calculation method for a servo motor of the present invention. The steps of the health warning calculation method of the servo motor include: first, obtaining an original vibration signal S100. Then, the vibration signal is sequentially transmitted to a data buffer S200. Next, the time-frequency domain conversion S300 is performed on the vibration signal. Wherein, the time-frequency domain conversion of the vibration signal can be Hilbert-Huang (Hilbert-Huang) Transform, HHT), Fast Fourier Transform (FFT), Wavelet Transform (WT) or other time-frequency domain conversion techniques. Finally, a deterioration index is calculated and a health index S400 is obtained. As for the detailed description of the health warning calculation method of the servo motor, please refer to the following.

Please refer to the fourth figure, which is a flow chart of the time-frequency domain conversion step of the health warning calculation method of the present invention. Taking Hilbert-Yellow Conversion (HHT) as an example, the step S300 is further described in detail, that is, the time-frequency domain conversion S300 is performed on the vibration signal. The vibration signal is read out through the data buffer, and then subjected to Empirical Mode Decomposition (EMD) decomposition to obtain a plurality of Intrinsic Mode Function (IMF) components S310. Then, the main inner model state function (IMF) component S320 is selected, and Hilbert-Huang transform (HHT) transform is performed on the main inner model state function (IMF) components to obtain a plurality of The signal S330 is instantaneously decomposed. Finally, the instantaneous decomposition signals are combined to obtain a Hilbert Spectrum S340. The Hilbert Spectrum is an amplitude-frequency-time three-dimensional map mainly describing the spectral content of the signal as a function of time so that the energy or intensity of the signal can be simultaneously represented in time and spectrum. .

Referring to FIG. 5A and FIG. 5B respectively, waveform diagrams of an original time domain vibration signal and a plurality of time domain decomposition signals and amplitude-frequency-time three-dimensional distribution diagrams of the time domain decomposition signals are respectively shown in the present invention. As shown in FIG. A, the original time domain vibration signal Sv can obtain a corresponding plurality of time domain decomposition signals St1~St9 through Hilbert-Huang transform (HHT), so as to make complex original time domain vibrations. The signal Sv is decomposed into a limited number of different time scales for signal resolution. In other words, these are the main (related) After the time domain decomposition signals St1 to St9 are superimposed, the original time domain vibration signal Sv can be restored to almost. In conjunction with the first figure B, the vibration detection of the servo motor 10 is taken as an example, and it is assumed that the rotation speed of the servo motor 10 is w(t). The original time domain vibration signal Sv returned by the vibration sensing unit 102 is transmitted through Hilbert-Huang transform (HHT) to obtain corresponding time domain decomposition signals St1 to St9. From bottom to top (see FIG. 5A), the first time domain decomposition signal St1 is 200 Hz, the second time domain decomposition signal St2 is 100 Hz, and the third time domain decomposition signal St3 is the servo motor 10 rotation speed w. The second harmonic of (t), the fourth time domain decomposition signal St4 is a multiple harmonic of w(t). As shown in FIG. 5B, the amplitude-frequency-time three-dimensional distribution map of the time domain decomposition signals St1 to St9 is the corresponding Hilbert Spectrum, and the heights in the figure show these The time domain resolves the strength (or energy) of the signals St1 to St9.

Please refer to the sixth figure, which is a schematic diagram of comparing the time domain decomposition signals of the present invention with a evaluation curve. The intensity of the time domain decomposition signals St1 to St9 at a certain instantaneous time is obtained according to the Hilbert Spectrum generated by the time frequency domain conversion unit 2062. Taking this embodiment as an example, nine instantaneous intensity values are extracted (circular black dots in the figure), and the instantaneous intensity values are compared with the evaluation curve to obtain a relative amount of intensity difference Δ G1~△g9, that is, the calculations of the intensity differences Δg1~Δg9 are the instantaneous intensity values at the corresponding frequencies minus the corresponding evaluation curve. The value of the evaluation curve can be regarded as a limit for measuring the health of the servo motor. As is apparent from the figure, the first intensity difference Δg1 and the second intensity difference Δg2 are both positive values, and also reflect the first time in the original vibration signal Sv at a frequency of 1,200 rps. The domain decomposition signal St1 and the vibration intensity of the second time domain decomposition signal St2 at a frequency of 800 rps are compared with the value of the evaluation curve It is the state of the deterioration (abnormal). In addition, the remaining intensity differences Δg3~Δg9 are all negative values, and similarly, the vibration intensity of the time domain decomposition signals St3~St9 in the original vibration signal Sv is compared with the value of the evaluation curve. It is normal (healthy) operating condition. It is worth mentioning that the degree of vibration deterioration of the servo motor is quantified by a deterioration index Di. Wherein, a maximum allowable value Tm is defined, and the calculation of the deterioration index Di is the sum of the intensity differences of the positive values (in the present example, the first intensity difference Δg1 and the second intensity difference Δg2). And the ratio of the maximum allowable value Tm. Among them, the maximum allowable value Tm can be obtained by the empirical rule of practical operation. That is, the deterioration index Di = (positive value intensity difference Δg1 ~ Δg9) / maximum allowable value Tm.

However, if the deterioration index Di exceeds 1, it is regarded as one. And, a health index Hi can be defined as: health index Hi = 1 - deterioration index Di.

It can be intuitively known that when the vibration of the servo motor 10 deteriorates to a lesser extent, the sum of the intensity differences Δg1 Δ Δg9 exceeding the value of the evaluation curve is larger, so that the calculated deterioration index Di is larger. In contrast, the health index Hi is also small.

In summary, the present invention has the following advantages:

1. With the built-in vibration sensing unit, the installation and wiring of the additional sensor can be omitted.

2. The vibration sensing unit (G-sensor) can be respectively disposed in the stator slot and the encoder of the servo motor, and the vibration can be separately performed by the corresponding driver and the servo motor. Estimation of vibration and transmission system vibration.

3. The health warning device of the servo motor can provide different health indicators of vibration and transmission system vibration in multiple directions.

However, the above description is only for the detailed description and the drawings of the preferred embodiments of the present invention, and the present invention is not limited thereto, and is not intended to limit the present invention. The scope of the patent application is intended to be included in the scope of the present invention, and any one skilled in the art can readily appreciate it in the field of the present invention. Variations or modifications may be covered by the patents in this case below.

10‧‧‧Servo motor

102‧‧‧Vibration sensing unit

20‧‧‧Servo drive

202‧‧‧High-speed serial communication interface

204‧‧‧Data buffer

206‧‧‧Microprocessor

2062‧‧‧time frequency domain conversion unit

2064‧‧‧ analytical unit

2066‧‧‧ Deterioration Index Calculation Unit

2068‧‧‧Health Index Calculation Unit

Claims (14)

A health warning device for a servo motor is applied to a vibration condition estimation of a computer numerical control tool machine; the health warning device of the servo motor comprises: a servo motor, which is built with at least one vibration sensing unit, Detecting an operating parameter of the servo motor and generating a vibration signal; and a servo driver connecting the servo motor, comprising a microprocessor, comprising a time frequency domain converting unit, receiving the vibration signal, and providing the vibration The signal is converted between the time domain and the frequency domain; an analysis unit is connected to the time-frequency domain conversion unit to receive the vibration signal and decompose the vibration signal into a plurality of decomposition signals; a deterioration index calculation unit is connected to the signal The parsing unit subtracts the selected decomposition signals from a evaluation curve to obtain a plurality of signal strength differences, and adds the signal strength differences of the positive values to calculate a deterioration index; and a health index a calculation unit, which is connected to the deterioration index calculation unit, and calculates a health index according to the deterioration index; thereby utilizing the built-in vibration The sensing unit can save the installation and wiring of the additional sensor, and according to the size of the health index, estimate the vibration condition of the servo motor running on the computer numerical control tool machine, and for nonlinearity Unsteady vibration characteristics provide better resolution. The invention relates to a health warning device for a servo motor according to claim 1, wherein the servo driver further comprises a high-speed serial communication interface for providing the vibration signal transmission. Communication interface. The invention relates to a health warning device for a servo motor according to claim 1, wherein the servo driver further comprises a data buffer connected to the high speed serial communication interface and the microprocessor to receive and store the vibration sensing unit. The vibration signal generated. For example, the health warning device for the servo motor of claim 1 is provided, wherein the vibration sensing unit is installed in an encoder of the servo motor to estimate the vibration state of the transmission system of the computer numerical control machine tool. The invention relates to a health warning device for a servo motor according to claim 1, wherein the vibration sensing unit is installed in a stator slot of the servo motor to estimate the vibration state of the machine of the computer numerical control tool machine. For example, the health warning device for the servo motor of claim 1 is wherein the data buffer is a buffer. For example, the health warning device for the servo motor of claim 1 is wherein the vibration sensing unit is a G-sensor. For example, the health warning device for the servo motor of the first application of the patent scope, wherein the evaluation curve is obtained by the rule of thumb. The health warning device for a servo motor according to claim 1, wherein the deterioration index is calculated by comparing the decomposition signals with the evaluation curve by the deterioration index calculation unit. A health warning calculation method for a servo motor is applied to a vibration condition estimation of a computer numerical control tool machine; the steps of the health warning calculation method of the servo motor include: (a) generating a vibration signal through a vibration sensing unit ; (b) sequentially transmitting the vibration signal to a data buffer; (c) performing a time-frequency domain conversion on the vibration signal through a time-frequency domain conversion unit; (c') receiving the vibration signal through a parsing unit and decomposing the vibration signal The signal is a plurality of decomposition signals; and (d) subtracting the decomposition signals from a evaluation curve by a deterioration index calculation unit to obtain a plurality of signal intensity differences, and summing the signal strength differences of the positive values, and calculating one Deteriorating the index and using a health index calculation unit to obtain a health index. For example, the health warning calculation method of claim 10, wherein the step (c) further comprises: (c1) performing empirical mode decomposition (EMD) to obtain a plurality of internal modeling state function (IMF) components; C2) selecting the main inner model state function (IMF) components; (c3) performing Hilbert-yellow transform (HHT) on the main inner model state function (IMF) components to obtain a plurality of Instantly decomposing the signals; and (c4) combining the instantaneous decomposition signals to obtain a Hilbert Spectrum. For example, the health warning calculation method of claim 10, wherein the time-frequency domain conversion of the vibration signal may be a Hilbert-Huang transform (HHT). For example, the health warning calculation method of claim 10, wherein the time-frequency domain conversion of the vibration signal can be a fast Fourier transform (FFT). For example, the health warning calculation method of claim 10, wherein the time-frequency domain conversion of the vibration signal may be a wavelet transform (WT).