JP3568939B2 - Method and apparatus for diagnosing state of rotating machine by analyzing shaft vibration - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method and an apparatus for predicting occurrence of a failure in advance by monitoring an amplitude value of a shaft vibration of a rotary machine such as a gas turbine.
[0002]
[Prior art]
A gas turbine is one of the typical rotating machines. The maintenance is very important together with the sale of the product, and a remote monitoring system using a public line has conventionally been adopted. For example, a maintenance company collectively monitors the operation data of gas turbines operating at delivery destinations in various locations, so that the status of gas turbines in various locations can be grasped from a distance, improving maintenance efficiency and increasing the number of operations. Of failure cases.
[0003]
Incidentally, since the gas turbine is a rotary machine, damage to the rotary bearing during long-time operation is an unavoidable problem. Once the failure occurs, repairing the gas turbine can be quite costly. In this case, by grasping the signs in advance, it is possible to minimize the damage due to the failure and reduce the repair cost. However, the signs are small and gradually appear in the operation data, and therefore, it is difficult for the maintenance staff to judge the operation data at a glance.
[0004]
Conventionally, a data collection device is installed near the rotating machine, and the shaft vibration data of the rotating machine is collected at a cycle of milliseconds or less, the collected data is frequency-analyzed, the frequency distribution is calculated, and the rotation is calculated in advance. The frequency distribution at the time of machine failure was compared with the frequency distribution calculated from the newly collected vibration data, and it was determined whether or not the newly collected vibration data was defective due to the degree of similarity between them. .
However, in this method, it is necessary to increase the data collection cycle for judging a defect, and as a result, the amount of data becomes enormous. Will increase. Further, the frequency analysis itself requires a large amount of calculation, and a load is required on the computer. Further, the frequency distribution includes noise and the distribution pattern is complicated, and specialized knowledge is required to determine a failure.
[0005]
For example, Japanese Patent Application Laid-Open No. H7-190849 discloses a technique for collecting shaft vibration data of a rotating machine and calculating a frequency distribution as a rotary bearing vibration diagnosis apparatus. A method is disclosed in which a frequency distribution due to disturbance such as noise is taught, and when a trouble occurs, the influence of the disturbance such as noise is removed by a neural network to judge a shaft trouble.
However, even in this technique, as in the above, it is necessary to reduce the data collection cycle for judging a problem, and the amount of data becomes enormous. Will increase. If there are a large number of rotating machines to be diagnosed, a diagnostic device is required for one machine. When diagnosing a large number of rotating machines with one diagnostic device, frequency analysis and calculation of a neural network impose a large load on the diagnostic device.
[0006]
[Problems to be solved by the invention]
The present invention has been made in view of the above points, and an object of the present invention is to use a shaft vibration data of a rotating machine to minimize a data amount necessary for a diagnosis in consideration of a remote diagnosis. By performing signal processing based on simple analysis and statistical calculations, it is possible to detect changes in state even from minute changes in shaft vibration, and easily and early determine signs of bearing damage. It is an object of the present invention to provide a method and apparatus for diagnosing a condition of a rotating machine by analyzing shaft vibration of the rotating machine which can reduce the economic burden and the like.
[0007]
[Means for Solving the Problems]
In order to achieve the above object, a method for diagnosing a condition of a rotating machine by shaft vibration analysis of the present invention collects vibration data of a shaft system of the rotating machine, transmits the collected vibration data to a data processing unit, and The standard deviation is calculated from the vibration data during operation, and the amount of change in the vibration data during operation is divided by a value obtained by multiplying the standard deviation by a coefficient to obtain defective shaft vibration data X, and | f (X) | The data X is processed by the formula f (X) such that amplification is performed when X |> 1 and attenuated when | X | <1 is detected, and amplified defective shaft vibration data is detected to determine a defective shaft vibration. It is configured.
[0008]
In the above method, when collecting vibration data, it is preferable that a representative value, for example, a maximum value is extracted from the vibration data within a certain period, and the vibration data necessary for diagnosis is reduced and used for vibration analysis. By reducing the amount of vibration data required for diagnosis and transmitting the data to the data processing unit, the communication load on a LAN (local area network), a public line, or the like can be reduced.
Further, in the above method, among the vibration data transmitted to the data processing unit, only the data during the operation of the rotating machine is extracted to perform the vibration analysis. In this case, for example, only the data at the time when the rotation speed of the rotating machine is larger than a certain value is extracted.
[0009]
A state diagnosis apparatus for shaft vibration analysis of a rotating machine according to the present invention is connected to a rotating machine, and collects operation data including rotation speed / vibration data of a shaft system, and data processing means for collecting the collected data. The standard deviation is calculated from the vibration data during normal operation using data communication means (public line, Internet, LAN, etc.) for transmitting data to the data collection means, and the amount of change in vibration data during operation. Is divided by a value obtained by multiplying a standard deviation by a coefficient to obtain defective shaft vibration data X. The equation f is such that amplification is performed when | f (X) | is | X |> 1, and is attenuated when | X | <1. (X) processing the data X and detecting amplified faulty shaft vibration data. The data processing means performs signal processing for determining a bearing damage sign. It is characterized in that a Shimesuru display means.
[0010]
In the above apparatus, a function of extracting a representative value, for example, a maximum value from vibration data within a certain period and reducing the amount of vibration data necessary for diagnosis is added to the data collection means, and a public line, a LAN, It is preferable to reduce such loads.
Further, in the above apparatus, the data processing means is provided with a function of extracting only data at a time when the rotation speed of the rotating machine is larger than a certain value, and using only data during operation for vibration analysis.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described. However, the present invention is not limited to the following embodiments, and can be implemented with appropriate modifications. FIG. 1 shows a system configuration for implementing a state diagnosis method based on shaft vibration analysis of a rotating machine according to a first embodiment of the present invention. The system includes a data collection unit 1 for collecting operation data including rotation speed / vibration data of a shaft system of a rotating machine, a data communication unit 2 for transmitting the collected data to a data processing unit 3 described below, and a data collection unit. The data processing unit 3 performs signal processing for determining a sign of bearing damage from the data transmitted from the unit 1, and the display unit 4 displays a result of the processing. A plurality of data collection units 1 may be prepared for diagnosing a plurality of rotating machines.
[0012]
FIG. 2 shows a specific system configuration example. From the data logger 10 and the control device 12 for collecting the operation data of the rotating machine such as the gas turbine to the small data processing device 14, the data logger 10 and the control device 12 operate using the general serial communication and analog data communication. Data collection is performed, and the small data processing device 14 performs preprocessing described later. Although not directly related to the present invention, the control device 12 controls not only data but also rotating machines. The operation data pre-processed by the small data processing device 14 (hereinafter, pre-process operation data) is transmitted to the data analysis (processing) device 18 via the public line 16 (or LAN or the like). The pre-processing operation data is analyzed by the data analyzer 18 and the result is displayed.
The data collection unit 1 in FIG. 1 corresponds to the data logger 10, the control device 12, and the small data processing device 14 in FIG. 2, and the data communication unit 2 in FIG. 1 corresponds to the public line 16 in FIG. The data processing unit 3 and the display unit 4 correspond to the data analysis device 18 in FIG.
[0013]
Data collection unit 1 of Figure 1, instead of transmitting all operational data to the data processing section 3, as pretreatment for transmission, as shown in FIG. 3, a period of time _{t n-1 ~t n, t} n .. Tn _{+ 1} ,... Are extracted from the operation data in each period according to a certain procedure such as maximum value extraction, and transmitted to the data processing unit 3 as preprocessing operation data. For example, in FIG. 3, data is transmitted to the data processing unit 3 at a sampling interval Δt, and P _n is transmitted as data at time t _n . In this example, _{P n} is the maximum value of the operating data _{t n-1} ~t _n period.
By this preprocessing, it becomes possible to generate a signal required by the data processing unit 3 in FIG. 1 while minimizing the load on the data communication unit 2. In FIG. 2, this preprocessing is performed by the small data processing device 14.
The data communication unit 2 in FIG. 1 can collect remote processing pre-processing operation data and a large number of pre-processing operation data in one place by using a public line or the Internet.
[0014]
The data processing unit 3 in FIG. 1 processes the pre-processing operation data transmitted from the data collection unit 1 and calculates a vibration diagnosis value.
First, since the preprocessing operation data includes both the data at the time of rotating machine operation and the data at the time of stoppage as shown in FIG. 4, only the data at the time when the rotation speed is larger than a certain value is extracted (hereinafter, extracted operation data).
The vibration diagnostic value is calculated using vibration data in the extracted operation data (hereinafter, extracted vibration data). The method will be described with reference to FIG. First, the average value m _p and the standard deviation σ _p of the extracted vibration data P _k−1 , P _k ,... Are calculated. Next, the difference ΔP between the value at a certain point in time and the value immediately before that is calculated according to Equation 1. k _p is a coefficient.
[0015]
ΔP = (P _k −P _k−1 ) / (k _p * σ _p ) Equation 1
( _Pk distribution is assumed to be normal distribution)
Next, the following processing on the difference value [Delta] P, and the vibration diagnostic value data H _k from Equation 2.
H _k = H _k−1 + f (ΔP) Equation 2
H ₀ = 0
Here, f (ΔP) is a monotonically increasing function satisfying the following condition.
| F (ΔP) | ≧ | ΔP | (when | ΔP | ≧ 1)
| F (ΔP) | <| ΔP | (when | ΔP | <1)
[0016]
By this process, the signal change amplified by f (ΔP) is accumulated with time. Whether there is a failure or not is determined by a maintenance person by automatically setting a threshold based on the vibration diagnosis value. The threshold value is determined based on an actual example and does not depend on the machine difference. 6 and 7 show two examples of calculation by the method of the present invention. In the calculation example, it is set to f (ΔP) = ΔP ^3. It can be seen that the vibration diagnosis value increases in both cases before the occurrence of the failure.
[0017]
Comparing the technology described in the embodiment of the present invention with the conventional technology described in the above-mentioned Japanese Patent Application Laid-Open No. H7-190849, it is necessary in the conventional technology to collect shaft vibration data at a sampling period of millisecond or less. On the other hand, in the present invention, the shaft vibration data may be collected at a sampling period of about one hour. Also, in the prior art, all the collected data is necessary for diagnosis, so that the communication amount becomes large. On the other hand, in the present invention, not all the collected data is necessary for the diagnosis, so the communication amount becomes small, The communication load when performing the operation is reduced. Furthermore, in the prior art, the diagnostic device performs Fourier transform calculation, noise spectrum reduction calculation, neural network calculation and complicated calculation, whereas in the present invention, only the analysis calculation including addition and multiplication and the statistical calculation are performed by the diagnostic device. This can be done, and the load on the computer is reduced.
[0018]
【The invention's effect】
The present invention is configured as described above, and has the following effects.
(1) It is possible to detect in advance a sign of a malfunction of the rotating machine from the shaft vibration.
(2) A state change can be detected even from a minute change in shaft vibration, and abnormality can be diagnosed at an early stage. Diagnosis is possible even with vibration data having a coarse sampling cycle.
(3) When remote diagnosis is performed on a rotating machine using a public line or the like, the data communication load can be reduced and the calculation load is also light, so that a plurality of rotating machines can be diagnosed with one diagnostic device. Further, the data transfer cost can be reduced.
(4) Since the determination of the occurrence of a failure can be made only by the degree of increase in the analysis value trend, a complicated determination criterion for the failure is unnecessary.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram showing a system configuration for implementing a state diagnosis method based on shaft vibration analysis of a rotating machine according to a first embodiment of the present invention.
FIG. 2 is a schematic configuration explanatory diagram illustrating an example of a system configuration according to the first embodiment of the present invention.
FIG. 3 is a graph for explaining signal processing (pre-processing) in a data collection unit according to the first embodiment of the present invention.
FIG. 4 is a graph for explaining a process of extracting data during driving according to the first embodiment of the present invention.
FIG. 5 is a graph for explaining signal processing in a data processing unit according to the first embodiment of the present invention.
FIG. 6 is a graph showing an example of calculating a vibration diagnostic value according to the method of the present invention.
FIG. 7 is a graph showing an example of calculating a vibration diagnostic value according to the method of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Data collection part 2 Data communication part 3 Data processing part 4 Display part 10 Data logger 12 Control device 14 Small data processing device 16 Public line 18 Data analysis device

Claims (10)

Collect vibration data of the shaft system of the rotating machine, send the collected vibration data to the data processing unit, calculate the standard deviation from the vibration data during normal operation, the amount of change in the vibration data during operation to the standard deviation By dividing by a value multiplied by a certain coefficient to obtain defective axis vibration data X, the equation f (X) amplifies if | f (X) | is | X |> 1, and attenuates if | X | <1. A method for diagnosing a state of a rotating machine by analyzing shaft vibrations, wherein the data X is processed, and amplified shaft vibration data is detected to determine a shaft vibration defect. 2. A state diagnosis by a shaft vibration analysis of a rotating machine according to claim 1, wherein a representative value is extracted from the vibration data within a certain period, and the vibration data necessary for the diagnosis is reduced and used for the vibration analysis. Method. 2. The state diagnosis method according to claim 1, wherein the method of collecting vibration data extracts a maximum value of vibration data within a certain period, reduces the amount of vibration data necessary for diagnosis, and uses the reduced amount of vibration data for vibration analysis. 4. The state diagnosis method according to claim 1, wherein the vibration analysis is performed by extracting only data during the operation of the rotary machine from the vibration data transmitted to the data processing unit. 5. The state diagnosis method according to claim 4, wherein only data at a time when the rotation speed of the rotating machine is larger than a predetermined value is extracted. A data collection unit connected to the rotating machine for collecting operation data including rotation speed and vibration data of the shaft system;
Data communication means for transmitting the collected data to the data processing means;
Using the data transmitted from the data collection means, calculate the standard deviation from the vibration data during normal operation, divide the amount of change in the vibration data during operation by a value obtained by multiplying the standard deviation by a certain coefficient, and X, and if | f (X) | is | X |> 1, the data X is processed by the formula f (X) to amplify and attenuate if | X | <1, and the amplified defective shaft vibration data is detected. Data processing means for performing signal processing for determining a bearing damage sign;
Display means for displaying a result processed by the data processing means;
A state diagnosis device based on shaft vibration analysis of a rotating machine, comprising: 7. The data collecting means according to claim 6, wherein a function of extracting a representative value from the vibration data within a certain period and reducing the amount of vibration data necessary for diagnosis is provided to reduce the load on the data communication means. A condition diagnosis device based on shaft vibration analysis of rotating machinery. 7. The rotating machine according to claim 6, wherein the data collection means is provided with a function of extracting the maximum value of the vibration data within a certain period and reducing the amount of vibration data necessary for diagnosis to reduce the load on the data communication means. Diagnosis device by shaft vibration analysis. 9. The rotation according to claim 6, wherein the data processing means is provided with a function of extracting only data at a time when the rotation speed of the rotating machine is larger than a certain value and using only data during operation for vibration analysis. A condition diagnostic device based on machine shaft vibration analysis. 10. The state diagnosis apparatus according to claim 6, wherein the data communication means is at least one of a public line, the Internet, and a LAN.

Images