JP2004361286A - Method of diagnosing deterioration of rotary machine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To remove vibration unrelated to deterioration of a facility to prevent an evaluation result of the deterioration from being fluctuated, and to indicate the optimum maintenance content thereby. <P>SOLUTION: This deterioration diagnostic method for a rotary machine is provided with a process for measuring the vibration generated in the rotary machine 11 by a vibration sensor 14, for extracting the vibration related to the deterioration, and for arranging amplitude representative values A of an amplitude distribution therein in order of the measurement to prepare the first data group, a process for finding a distribution index DA of the second data group constituting by extracting, from the first data group, a prescribed number of the amplitude representative values A containing the up-to-date amplitude representative values A and successive in the measured order, and a process for measuring the vibration of the rotary machine 11 to find amplitude representative values B of an amplitude distribution therein, so as to be added to the second data group, and for constituting the new second data group by removing the oldest amplitude representative value A from the second data group to find a distribution index DB thereof. The method includes also a process for adding the amplitude representative values B to the first data group, when the distribution indexes DA, DB are substantially consistent each other, and for deleting the amplitude representative values B when not consistent substantially. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method of diagnosing deterioration of a rotating machine, and more particularly, to evaluating a deterioration state of a rotating machine using an output signal of a vibration detection sensor attached to the rotating machine, an output specific to machine equipment unrelated to deterioration. The present invention relates to a method for diagnosing deterioration of a rotating machine that removes a signal to prevent a change in an evaluation result.
[0002]
[Prior art]
Conventionally, for example, in a method of diagnosing an abnormality of a low-speed rotating machine such as a roller apron of a continuous casting machine, a signal detected by a vibration sensor is applied to a band-pass filter, and then envelope processing is performed. The difference between the upper limit value and the lower limit value is calculated, and when this difference exceeds a predetermined threshold value, it is determined that there is an abnormality (for example, see Patent Document 1).
In addition, a diagnostic device used for diagnosing deterioration of a rotating machine continuously extracts output signals from the rotating machine for several seconds by a vibration detection sensor and performs various processes, for example, dividing a vibration frequency into a plurality of frequency regions. Then, an index value (hereinafter, simply referred to as a vibration value) indicating a vibration amplitude value or a probability density distribution of a vibration waveform is obtained for each frequency region, and a preset reference value of the vibration value for each deterioration level is set. The deterioration level was determined in comparison with the above.
Then, when actually performing the deterioration diagnosis of the rotating machine, the deterioration diagnosis is performed using only the vibration value in order to emphasize the versatility and efficiency of the deterioration diagnosis, and the same is applied to all the deterioration types generated in the rotating machine. It was common to make a judgment with a reference value.
[0003]
[Patent Document 1]
JP-A-7-270228
[0004]
[Problems to be solved by the invention]
In general, even if a rotating machine is used under constant operating conditions such as load and rotation speed, the vibration value obtained during a time or period (for example, on a daily basis) in which the deterioration state of the rotating machine can be considered constant. Fluctuates remarkably over time, as shown in FIG.
For this reason, when the deterioration progresses to some extent and the vibration value approaches the judgment reference value, the influence of the signal peculiar to the mechanical equipment irrelevant to the deterioration cannot be ignored, and a large fluctuation occurs in the required vibration value, and the degree of measurement is increased. Each time, the abnormality determination and the normality determination are repeated. As a result, the efficiency of the deterioration determination work has been reduced, and there has been a problem in the accuracy of the deterioration determination. In addition, it is difficult to detect deterioration at an early stage, which makes it difficult to cope with equipment maintenance. Further, in a rotating machine that performs intermittent operation, a so-called unsteady operation state before the stationary operation state area is reached frequently occurs, so that the determination result of the deterioration diagnosis may further fluctuate.
[0005]
In particular, when performing a deterioration diagnosis of a rotating machine, since the judgment is performed using the same reference value for all the deterioration types, the deterioration levels indicated by the judgment reference values are set extremely coarsely, and the deterioration is determined. If it is determined that the deterioration has occurred, it is necessary to separately perform a detailed diagnosis of the deterioration to specify the type of the deterioration. Then, after the type of deterioration is identified by the detailed diagnosis, based on the time change of the vibration value, for example, the transition of the vibration value simply averaged in units of 24 hours, the person can determine the action content for equipment maintenance. Had been decided.
Here, as a system for performing a precise diagnosis of deterioration, for example, a frequency analysis of a signal detected by a vibration detection sensor is performed to specify a type and an abnormal level of an abnormality, and comments on whether or not repair is performed and when the repair is performed. A device having a function of outputting has been proposed. However, in this system, even if the necessity of repair can be determined, effective maintenance contents cannot be instructed according to the type of deterioration.For example, when an abnormal state can be resolved simply by simple measures such as refueling Even in this case, there is a case where a meaningless and inefficient countermeasure such as a disassembly check requiring a large amount of cost and a lot of time is selected.
[0006]
The present invention has been made in view of the above circumstances, to prevent the evaluation result of the deterioration state from largely fluctuating by removing a signal peculiar to mechanical equipment irrelevant to deterioration, and to specify a deterioration type and a deterioration level. It is an object of the present invention to provide a method of diagnosing deterioration of a rotating machine capable of instructing an optimum maintenance content corresponding thereto.
[0007]
[Means for Solving the Problems]
A method for diagnosing deterioration of a rotating machine according to the present invention, which meets the above-described object, comprises measuring vibration generated in the rotating machine for a predetermined time at predetermined time intervals using a vibration detection sensor attached to the rotating machine. Extracting vibrations related to mechanical deterioration, obtaining respective amplitude distributions R, and arranging amplitude representative values A defining the respective amplitude distributions R in a measurement order to create a first data group;
A distribution index DA of a second data group is obtained by extracting a predetermined number of amplitude representative values A including the latest amplitude representative value A from the amplitude representative value A of the first data group and including the latest amplitude representative value A in measurement order. Process and
The vibration generated by the rotating machine is measured for a predetermined time at predetermined time intervals using the vibration detection sensor, and an amplitude representative value B that defines the amplitude distribution S is obtained from the amplitude distribution S, and the second And forming a new second data group by removing the oldest amplitude representative value A from the second data group and obtaining its distribution index DB;
When the distribution index DA of the second data group substantially matches the distribution index DB of the new second data group, the amplitude representative value B is set as the latest amplitude representative value A in the first data group. A step of deleting the amplitude representative value B when the distribution index DA of the second data group and the distribution index DB of the new second data group do not substantially match.
[0008]
When deterioration occurs in a rotating machine, vibration due to the deterioration (when the frequency range of the vibration is an audible sound range is recognized as an abnormal sound), the magnitude of the deteriorated vibration (vibration amplitude) progresses. Increase with.
For this reason, if the vibration amplitude of the vibration generated during the operation of the rotating machine is detected for a predetermined time (for example, 1 to 3 seconds) at predetermined time intervals (for example, 1 to 3 minutes), the vibration The occurrence and progress of deterioration can be detected from the amplitude change process.
Here, the amplitude distribution R is obtained from the vibration amplitude of the vibration related to the detected deterioration, and an amplitude representative value A that quantitatively represents the characteristics of the amplitude distribution R is obtained to constitute the first data group. The number of data can be reduced efficiently. Furthermore, since the first data group is configured by arranging the amplitude representative values A in the order of measurement, the temporal progress of deterioration can be easily grasped using the first data group.
[0009]
In addition, the vibration detection sensor detects vibration caused by deterioration and also detects vibration peculiar to the rotating machine unrelated to deterioration.
Therefore, a predetermined number (for example, 100) of amplitude representative values A including the latest amplitude representative value A and consecutive in the measurement order is extracted from the amplitude representative value A of the first data group to form the second data group. Then, the distribution index DA is obtained. By obtaining the distribution index DA, the characteristics of the distribution of the second data group can be quantitatively grasped.
Next, vibration generated by the rotating machine is measured for a predetermined time (for example, 1 to 3 seconds) at predetermined time intervals (for example, 1 to 3 minutes) using a vibration detection sensor, and the amplitude distribution S is used to calculate the amplitude. An amplitude representative value B that defines the distribution S is obtained, the amplitude representative value B is added to the second data group, and a new second data group is formed by removing the oldest amplitude representative value A from the second data group. And its distribution index DB. Thereby, the characteristics of the distribution of the new second data group can be grasped quantitatively.
[0010]
Here, when the amplitude representative value B is obtained from the vibration caused by the deterioration, the distribution of the new second data group and the distribution of the second data group are considered to be substantially the same. When the amplitude representative value B is obtained from vibration peculiar to a rotating machine unrelated to deterioration, the distribution of the new second data group and the distribution of the second data group are considered to be different distributions. .
For this reason, when the distribution index DA and the distribution index DB substantially match, the amplitude representative value B is regarded as being obtained from the vibration caused by the deterioration. The amplitude representative value A is added to the first data group.
On the other hand, if the distribution index DA and the distribution index DB do not match, the amplitude representative value B is regarded as being obtained from the vibration peculiar to the rotating machine unrelated to the deterioration. Not added to the first data group.
[0011]
In the method for diagnosing deterioration of a rotating machine according to the present invention, it is preferable that the amplitude distributions R and S are obtained by dividing the vibration into a plurality of frequency regions and for each of the frequency regions.
Since a rotating machine is made up of various mechanical components, when deterioration occurs in each component, the generated vibration exists in a specific frequency region corresponding to each component.
Therefore, if the frequency domain is set according to each component and the amplitude distribution of the vibration is obtained for each frequency domain, it is possible to determine which component is deteriorating and progressing, and to determine the type of degradation. Can be identified.
[0012]
In the method for diagnosing deterioration of a rotating machine according to the present invention, the average values of the amplitude distributions R and S are used as the amplitude representative values A and B, and the average values of the average values are used as the distribution indexes DA and DB. Process capability index C indicating the degree of variation _pi It is preferable to employ
When no deterioration occurs in each component of the rotating machine, or when the deterioration does not progress even if the deterioration occurs, the vibration generated from the rotating machine is generated at predetermined time intervals (for example, 1 to 3 minute intervals). Each of the amplitude distributions R and S obtained by measuring only a predetermined time (for example, 1 to 3 seconds) can be approximated as a normal distribution.
Therefore, as the representative amplitude values A and B that quantitatively represent the characteristics of the amplitude distributions R and S, the average value of the amplitude distributions R and S at which the probability of being detected when the vibration is measured is maximized. It is preferable to employ it.
[0013]
When the second data group (new second data group) is formed, each amplitude representative value A (each amplitude representative value B) has the shape of the distribution of the second data group (new second data group). Reflects between the upper and lower limits.
Then, a process capability index C indicating the degree of variation of the average value of each amplitude distribution R, S as each distribution index DA, DB. _pi , The relationship between the difference between the upper limit value and the lower limit value (also referred to as a standard width) and the “variation” can be quantitatively expressed.
Here, the process capability index C indicating the degree of variation of the average value _pi Can be calculated by the following equation, where T is the difference between the upper limit value and the lower limit value, and σ is the standard deviation of the second data group (new second data group).
C _pi = T / (6σ) (1)
[0014]
In the method for diagnosing deterioration of a rotating machine according to the present invention, a predetermined number of the representative amplitude values A successively extracted from the first data group are sequentially extracted to form a plurality of third data groups, and each of the third data groups is formed. The representative value and the statistical index of the distribution of each third data group are sequentially obtained for each group, the deterioration level is determined from the magnitude of each representative value, and the change rate of the statistical index of each third data group is determined. It is preferable to determine the maintenance content of the rotary machine from the following.
[0015]
A plurality of third data groups arranged in chronological order from the first data group are formed, a representative value is obtained for each of the third data groups, and the deterioration level is indicated by the representative value. The change behavior can be efficiently reduced and the change behavior can be clearly shown.
Further, by obtaining a statistical index of distribution for each third data group, for example, it is possible to quantitatively grasp the degree of bias of the distribution of each third data group. Therefore, if a statistical index is selected such that the statistical index is small when the degree of bias of the distribution of the third data group is small, if the statistical index increases, each of the third data group It can be considered that the representative value is gradually increasing, and it can be determined that the deterioration is progressing. For this reason, the progress speed of deterioration can be grasped from the change speed of the statistical index, and an optimal maintenance action guideline can be issued according to the progress speed of deterioration.
[0016]
In the method for diagnosing deterioration of a rotating machine according to the present invention, a median M of each of the third data groups is adopted as each of the representative values, and an average value of the third data group is used as the statistical index. Process capability index C considering bias _pik It is preferable to employ
When deterioration occurs or progresses, each amplitude representative value A constituting each third data group tends to increase with time, and each amplitude representative value A is symmetric between the upper limit and the lower limit. Will not be distributed.
For this reason, a central value M determined by the upper limit and the lower limit is adopted as the representative value of each third data group. Thus, the difference between the third data groups can be clearly displayed using the representative values.
[0017]
That each amplitude representative value A constituting each third data group is not symmetrically distributed between the upper limit value and the lower limit value is equivalent to a difference between the median value of the distribution and the average value of the distribution. .
Therefore, as a statistical index, the process capability index C considering the bias of the average value is used. _pik By using, the bias between the median and the average in each third data group can be quantitatively grasped.
Here, the process capability index C considering the bias of the average value _pik Can be calculated by the following equation, where SU and SL are the upper and lower limits, respectively, and the average and standard deviation of the third data group are μ and s.
C _pik = MIN {(SU-μ) / (3s), (μ-SL) / (3s)} (2)
Here, MIN {...} Is a function for selecting the minimum value in parentheses.
[0018]
In the method for diagnosing deterioration of a rotary machine according to the present invention, the process capability index C _pik It is preferable to set an upper limit value and a lower limit value for each of the median values M in advance when calculating.
Although the distribution shape of the third data group changes with the progress of deterioration, the change differs depending on the type of deterioration. For example, there are a case where the distribution shape tends to spread and a case where the distribution shape tends to narrow as the deterioration progresses.
Therefore, by setting an upper limit value and a lower limit value for the magnitude of the median in advance for each type of deterioration, a process in which the bias of the average value of the third data group is considered in accordance with the progress of deterioration. Capacity index C _pik Can be determined exactly.
[0019]
In the method for diagnosing deterioration of a rotary machine according to the present invention, a frequency range to which the median M belongs and a process capability index C corresponding to the median M for each of the medians M. _pik Find the rate of change of
(1) Correspondence between each frequency region and the degradation type, (2) Correspondence between each degradation level included in each degradation type and the range of the median value, (3) Included in each degradation type set in advance Each maintenance action guideline and process capability index C within each deterioration level _pik By comparing with the corresponding relationship of the range of change speed of
The deterioration type and the deterioration level can be specified, and the specific maintenance content can be determined.
[0020]
As the deterioration progresses, the bias of the distribution of the third data group increases, and the process capability index C _pik Increase. Therefore, the process capability index C _pik The rate at which the deterioration proceeds can be ascertained from the rate of change.
As a result, (1) the correspondence relationship between each frequency region and the degradation type, (2) the correspondence relationship between each degradation level included in each degradation type and the range of the median value, and (3) the degradation level included in each degradation type. Action Guidelines and Process Capability Index C _pik By setting in advance the correspondence of the range of the change speed of each, the optimum maintenance action guideline according to the current situation can be obtained by grasping the deterioration level and the progression speed of the specified deterioration type. Can be easily obtained.
[0021]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, embodiments of the present invention will be described with reference to the accompanying drawings to provide an understanding of the present invention.
Here, FIG. 1 is an explanatory diagram of a rotating machine deterioration diagnosis apparatus to which a rotating machine deterioration diagnosis method according to an embodiment of the present invention is applied, and FIG. 2 is a diagram illustrating a vibration detection sensor including a signal unique to the rotating machine. FIG. 3 is an explanatory diagram showing an upper limit value and a lower limit value set respectively with respect to a temporal change of a median value of two deterioration types, and FIG. 4 is a process capability index C. _pik FIG. 6 is an explanatory diagram showing a time change of the scalar.
As shown in FIG. 1, a rotating machine deterioration diagnosis apparatus 10 to which a rotating machine deterioration diagnosis method according to one embodiment of the present invention is applied detects vibrations generated in a rotating machine 11 and data for deterioration diagnosis. And a deterioration diagnosis system 13 for performing a deterioration diagnosis based on the obtained deterioration diagnosis data. Hereinafter, these will be described in detail.
[0022]
The data creation system 12 includes a vibration detection sensor 14 that measures the vibration of the rotating machine 11 at predetermined time intervals (for example, 1 to 3 minutes) for a predetermined time (for example, 1 to 3 seconds), and a vibration detection sensor 14. An amplifier 15 for amplifying an output signal from the amplifier, a band-pass filter 16 for dividing the amplified signal into a plurality of frequency regions, and obtaining amplitude distributions R and S from the respective signals divided for each frequency region. An average value output unit 17 having a function of calculating and outputting an average value of amplitude distributions R and S, which are examples of amplitude representative values A and B defining R and S, is provided.
Further, the data creation system 12 includes a first data group input / output unit 18, a first process capability index calculating unit 19, another first process capability index calculating unit 20, and the same first process capability index. Sex determining means 21 is provided.
Here, the first data group input / output means 18 receives the average value as the amplitude representative value A output from the average value output means 17 based on a command from the outside and arranges the first data group by arranging in order of measurement. It has a function of creating and outputting the created first data group.
[0023]
The first process capability index calculating means 19 calculates a predetermined number (for example, 100) of the amplitude representative values A including the average value as the latest amplitude representative value A and continuing in the measurement order from the first data group input / output means 18. , A second data group is formed by extracting the average value as a process capability index C indicating the degree of variation of the average value, which is an example of the distribution index DA of the second data group. _pi Is provided.
Further, another first process capability index calculating unit 20 calculates the average value as the amplitude representative value B output from the average value output unit 17 and the second data generated by the first process capability index calculating unit 19. The group is accepted, the average value as the amplitude representative value A which is the oldest in the second data group is deleted to form a new second data group, and the average value which is an example of the distribution index DB of the new second data group Capability index C indicating the degree of variation of _pi Is provided.
[0024]
Further, the first process capability index identity determining means 21 calculates the process capability indices C calculated by the first process capability index calculating means 19 and 20 respectively. _pi Of each process capability index C _pi Are substantially the same, the average value as the amplitude representative value B output from the average value output means 17 is output to the first data group input / output means 18 as the latest average value as the amplitude representative value A. And each process capability index C _pi Has a function of deleting the average value as the amplitude representative value B output from the average value output means 17 when the values do not substantially coincide with each other.
[0025]
With such a configuration, when performing the deterioration diagnosis of the rotating machine 11, the initial data as the first data group can be created from the measured amplitude distribution R of the vibration. Further, it is determined whether or not the measured amplitude distribution S of the vibration is the vibration caused by the deterioration, and only the average value as the amplitude representative value B of the vibration caused by the deterioration is sequentially added to the first data group. Can be added.
Here, those conventionally used can be applied to the vibration detection sensor 14, the amplifier 15, and the band-pass filter 16. The average value output unit 17, the first data group input / output unit 18, the first process capability index calculation units 19 and 20, and the first process capability index identity determination unit 21 are provided with the functions described above. Can be configured by, for example, mounting a program that expresses in a personal computer.
[0026]
The deterioration diagnosis system 13 has a function of sequentially extracting an average value of a predetermined number (for example, 200) of amplitudes from the first data group and forming a plurality of third data groups. A data group creating unit 22; a median calculating unit 23 having a function of calculating a median M which is an example of a representative value of each third data group for each third data group; Process capability index C taking account of the bias of the average value, which is an example of the statistical index of the distribution of the data group of FIG. _pik And a second process capability index change speed calculating means 24 having a function of sequentially calculating the change speed and calculating the change speed.
Further, the deterioration diagnosis system 13 includes a correspondence relation between the deterioration type generated in the rotating machine 11 and a frequency region of the generated vibration, a correspondence relation between the deterioration level and the median value range included for each deterioration type, and the deterioration relation included in each deterioration type. Action guidelines and process capability index C within the specified deterioration level _pik It has a deterioration determination and maintenance action database 25 having a function of creating and storing the correspondence of the change speed ranges in advance, a deterioration determination means 26, and a maintenance action guideline determination means 27.
[0027]
Here, the deterioration determination unit 26 compares the median M sequentially output from the median calculation unit 23 and the frequency region to which the median M belongs with each frequency region stored in the deterioration determination and maintenance action database 25 and the deterioration type. A function is provided for determining the presence / absence of deterioration, the type of deterioration, and the level of deterioration by comparing the corresponding relationship and the corresponding relationship between each deterioration level and the range of the median included in each deterioration type.
Further, the maintenance action guideline determination means 27 determines the process capability index C corresponding to the median M sequentially output from the second process capability index change speed calculating means 24. _pik The change rate of the deterioration is stored in the deterioration judgment and maintenance action database 25. Each maintenance action guideline and process capability index C within each deterioration level included in each deterioration type are shown. _pik It has a function to output a corresponding maintenance action guideline in comparison with the corresponding relationship of the range of the change speed of.
Further, the deterioration diagnosis system 13 has a display means 28 for displaying the median M sequentially output from the median calculation means 23, the judgment result of the deterioration judgment means 26, and the judgment result of the maintenance action guideline judgment means 27. ing.
[0028]
With such a configuration, a plurality of third data groups arranged in a time series from the first data group are formed, and the median M of the distribution is obtained for each group. It is possible to clearly show the occurrence of deterioration and the change behavior thereof by reducing the amount of deterioration. Then, the type of deterioration can be grasped from the frequency region where the median M exists, and the degradation level can be grasped from the magnitude of the median M.
Also, the process capability index C corresponding to the median M _pik It is possible to grasp the progress speed of the deterioration from the change speed of, and it is possible to easily obtain the optimal maintenance action guideline according to the deterioration type and the deterioration level.
Here, the third data group creation means 22, the median value calculation means 23, the second process capability index change speed calculation means 24, the deterioration judgment and maintenance action database 25, the deterioration judgment means 26, and the maintenance action guideline judgment means 27 can be configured by installing a program that expresses the above-described functions in, for example, a personal computer. Further, as the display means 28, various display devices for a personal computer which are commercially available can be used.
[0029]
Next, a method for diagnosing deterioration of a rotating machine according to one embodiment of the present invention will be described in detail.
The deterioration diagnosis of the rotating machine 11, for example, the blower in which the fan was directly attached to the output shaft of the motor was performed.
First, vibration detection sensors 14 (for example, acceleration sensors) are attached to bearings on both sides of the motor, and the vibration of the rotary machine 11 is set to a predetermined time (for example, 1 to 3 minutes) at predetermined time intervals (for example, 1 to 3 minutes). 3 seconds). After the output signal from the vibration detection sensor 14 was amplified by the amplifier 15, it was divided by the bandpass filter 16 into a plurality of frequency regions (for example, 1 to 3 kHz, 3 to 40 kHz, and 10 to 40 kHz). Here, the signal (P, Q, R) specific to the rotating machine is suddenly mixed into the signal output from the vibration detection sensor 14, as shown in FIG. For this reason, each of the divided signals is input to the average value output means 17 to extract a vibration related to the deterioration of the rotary machine 11 and obtain an amplitude distribution R of the vibration, and calculate the average value as the amplitude representative value A of the amplitude distribution R. The first data group is calculated and sequentially input to the first data group input / output unit 18 to create initial data (for example, 100 to 500 data pieces) in the first data group arranged in the measurement order.
[0030]
After the creation of the initial data in the first data group was completed, the deterioration diagnosis of the rotating machine 11 was started.
The first process capability index calculating means 19 extracts a predetermined number (for example, 100) of average values including the latest average value from the first data group input / output means 18 and including the latest average value, and outputs the second data. And a process capability index C as a distribution index DA of the second data group _pi Is calculated.
Next, the first process capability index calculating unit 20 calculates the average value as the amplitude representative value B output from the average value output unit 17 and the second data group created by the first process capability index calculating unit 19. Accepting, the oldest average value in the second data group is deleted to form a new second data group, and the process capability index C as the distribution index DB of the new second data group _pi Is calculated.
[0031]
Subsequently, each process capability index C calculated by each of the first process capability index calculation units 19 and 20 by the first process capability index identity determination unit 21. _pi Of each process capability index C _pi Are substantially the same, the average value as the representative amplitude value B output from the average value output means 17 is output to the first data group input / output means 18 as the average value as the latest amplitude representative value A. As a result, the latest average value is added to the first data group, and the number of data increases by one.
On the other hand, each process capability index C _pi Do not substantially coincide with each other, the average value as the amplitude representative value B output from the average value output means 17 is deleted. As a result, the number of data does not change in the first data group.
Here, each process capability index C _pi Is, for example, 0.25 or less, each process capability index C _pi Are determined to be substantially the same, and each process capability index C _pi Is greater than 0.25, for example, each process capability index C _pi Are determined not to substantially match.
In addition, each process capability index C _pi Is determined by each process capability index C obtained for each frequency range of the measured vibration. _pi For each process capability index C in all frequency ranges. _pi Each average value is added to the first data group only when the identity of is confirmed.
[0032]
The first data group stored in the first data group input / output means 18 is read by the third data group creation means 22, and each average value arranged in the measurement order is determined for each frequency domain. A plurality of third data groups are formed by dividing the data by the number (for example, 200). Then, the median value calculating means 23 calculates the median value M for each third data group for each frequency region.
Further, the second process capability index change speed calculating means 24 calculates the process capability index C for each third data group for each frequency region. _pik Are sequentially obtained, and the change speed is calculated.
Here, the process capability index C _pik Is calculated based on the above-described equation (2) using an upper limit value and a lower limit value preset for the median M of each third data group for each frequency region. FIG. 3 shows the relationship between the upper limit value and the lower limit value set for the median value M that changes with time in the two types of deterioration.
[0033]
The deterioration determining means 26 determines the correspondence between each frequency region and the deterioration type stored in the deterioration determination and maintenance action database 25 and the correspondence between each deterioration level and the median range included in each deterioration type. The deterioration type and the deterioration level are read and compared with the median value M for each frequency domain sequentially output from the median value calculation means 23 and displayed on the display means 28.
The maintenance action guideline determining means 27 calculates the process capability index C for each frequency region calculated by the second process capability index change speed calculating unit 24. _pik With the process capability index C as shown in FIG. _pik Of the maintenance action guidelines and the process capability index C within the respective degradation levels included in the respective degradation types stored in the degradation determination and maintenance action database 25 as well as the time change of the _pik Is read, and the process capability index C for each frequency region sequentially output from the second process capability index change speed calculating means 24 is read. _pik The corresponding maintenance action guideline is determined by comparing with the change speed of, and the content is displayed on the display means 28.
Here, the maintenance action guideline indicates, for example, facility management guidance such as "continue regular cycle monitoring", "perform monitoring every short cycle", "return to regular cycle monitoring", "plan an inspection plan", and "stop operation immediately".
[0034]
As described above, the embodiments of the present invention have been described. However, the present invention is not limited to these embodiments, and can be changed without changing the gist of the invention. A case where the method for diagnosing deterioration of a rotating machine of the present invention is configured by combining some or all of the forms and modifications is also included in the scope of the present invention. For example, the measurement of the vibration of the rotating machine is performed at intervals of 1 to 3 minutes and for a period of 1 to 3 seconds. Can be changed to
Although the division by the band-pass filter is made into three divisions of 1 to 3 kHz, 3 to 40 kHz, and 10 to 40 kHz, the width of the frequency domain and the number of divisions can be arbitrarily set according to the type of the component of the rotating machine.
Although the number of data in the second data group is set to 100, the number of data may be set to less than 100 or to 101 or more according to the target of the rotating machine for performing the deterioration diagnosis. In addition, the number of data in the third data group is set to 200, but the number of data may be set to less than 200 or to 201 or more according to the target of the rotating machine for performing the deterioration diagnosis.
[0035]
【The invention's effect】
In the method for diagnosing deterioration of a rotating machine according to any one of claims 1 to 7, vibration generated in the rotating machine is measured for a predetermined time at predetermined time intervals using a vibration detection sensor attached to the rotating machine. Extracting vibrations related to deterioration, obtaining respective amplitude distributions R, and arranging amplitude representative values A defining the respective amplitude distributions R in the order of measurement to create a first data group; A step of obtaining a distribution index DA of a second data group constituted by extracting a predetermined number of amplitude representative values A including the latest amplitude representative value A from the value A and being continuous in the measurement order; The amplitude representative value B that defines the amplitude distribution S is determined from the amplitude distribution S by measuring the amplitude distribution S for a predetermined time using a vibration detection sensor, and is added to the second data group. Oldest from A step of forming a new second data group except for the amplitude representative value A and obtaining its distribution index DB, and the distribution index DA of the second data group and the distribution index DB of the new second data group are substantially Is added to the first data group as the latest amplitude representative value A, the distribution index DA of the second data group and the distribution index DB of the new second data group substantially match. If not, a step of deleting the amplitude representative value B is provided, so that when measuring the vibration of the rotating machine, it is possible to remove a signal peculiar to mechanical equipment unrelated to deterioration.
As a result, even if the deterioration progresses to some extent and approaches the determination reference value, the deterioration evaluation result does not change and stable deterioration determination can be performed.
[0036]
In particular, in the method for diagnosing deterioration of a rotating machine according to the second aspect, since each of the amplitude distributions R and S divides the vibration into a plurality of frequency regions and obtains each for each frequency region, it is easy to specify the type of deterioration. , It is possible to determine in advance the content of the maintenance response to the deterioration.
[0037]
In the method for diagnosing deterioration of a rotating machine according to claim 3, the average values of the amplitude distributions R and S are used for the amplitude representative values A and B, and the degree of variation of the average values is used for the distribution indexes DA and DB. Process capability index C _pi Is employed, it is possible to quantitatively evaluate the identity of the second data group and the new second data group, and to reliably remove signals specific to mechanical equipment unrelated to deterioration. Become.
[0038]
In the method for diagnosing deterioration of a rotating machine according to claim 4, a predetermined number of successive amplitude representative values A are sequentially extracted from the first data group to form a plurality of third data groups, and each of the third data groups is formed. For each time, the representative value and the statistical index of the distribution of each third data group are sequentially obtained, the level of deterioration is determined from the magnitude of each representative value, and the rotational speed of the rotating machine is determined from the rate of change of the statistical index of each third data group. Since the maintenance content is determined, it is possible to grasp the occurrence and progress of the deterioration one by one and to take an early response to the deterioration.
[0039]
In the method for diagnosing deterioration of a rotating machine according to the fifth aspect, the median value M of each third data group is adopted as each representative value, and the bias of the average value of the third data group is used as a statistical index. Process capability index C considered _pik Is adopted, it is possible to clarify the difference between the respective third data groups and quantitatively display the deterioration level.
In addition, the statistical index includes a process capability index C in consideration of the bias of the average value of the third data group. _pik Is employed, the bias of the distribution of the median value and the average value can be quantitatively grasped, and the progress of deterioration can be quantitatively evaluated.
[0040]
In the method for diagnosing deterioration of a rotary machine according to claim 6, the process capability index C _pik Is calculated in advance, the upper limit value and the lower limit value are set in advance for each median M, so that the process capability index C considering the deviation of the average value of the third data group in accordance with the progress of deterioration _pik Can be accurately obtained, and the progress of deterioration can be more accurately evaluated.
[0041]
In the method for diagnosing deterioration of a rotating machine according to claim 7, a frequency range to which the median M belongs for each median M, and a process capability index C corresponding to the median M. _pik (1) Correspondence relationship between each frequency region and degradation type, (2) Correspondence relationship between each degradation level included in each degradation type and the range of median value, (3) Each maintenance action guideline and process capability index C within each deterioration level included in each deterioration type _pik Since the type of deterioration and the level of deterioration are specified and specific maintenance contents are determined by comparing with the correspondence relationship of the range of the change speed, a sensitive response is obtained even if the amplitude representative value A fluctuates sharply with time. In addition, the progress of deterioration can be reliably evaluated, and the progress of deterioration can be recognized at an early stage.
Furthermore, since specific maintenance action guidelines can be obtained in accordance with the type of deterioration, deterioration level, and progress of deterioration, it is possible to carry out effective equipment management without having specialized knowledge. Become.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of a rotating machine deterioration diagnosis apparatus to which a rotating machine deterioration diagnosis method according to an embodiment of the present invention is applied.
FIG. 2 is an explanatory diagram showing an output signal from a vibration detection sensor including a signal unique to a rotating machine.
FIG. 3 is an explanatory diagram showing a relationship between an upper limit value and a lower limit value set with respect to a time-varying median value M in two types of deterioration.
FIG. 4 Process capability index C _pik FIG. 6 is an explanatory diagram showing a time change of the scalar.
FIG. 5 is an explanatory diagram showing an output example of a vibration value in a conventional example.
[Explanation of symbols]
10: Deterioration diagnosis device for rotating machine, 11: Rotating machine, 12: Data creation system, 13: Deterioration diagnosis system, 14: Vibration detection sensor, 15: Amplifier, 16: Band pass filter, 17: Average value output means, 18 : First data group input / output means, 19, 20: first process capability index calculation means, 21: first process capability index identity determination means, 22: third data group creation means, 23: center Value calculation means, 24: change rate calculation means for the second process capability index, 25: deterioration judgment and maintenance action database, 26: deterioration judgment means, 27: maintenance action guideline judgment means, 28: display means

Claims (7)

Vibration generated by the rotating machine is measured for a predetermined time at predetermined time intervals using a vibration detection sensor attached to the rotating machine, and vibration associated with the deterioration of the rotating machine is extracted, and its amplitude distribution R is calculated. Determining and arranging the amplitude representative values A defining the respective amplitude distributions R in the order of measurement to create a first data group;
A distribution index DA of a second data group is obtained by extracting a predetermined number of amplitude representative values A including the latest amplitude representative value A from the amplitude representative value A of the first data group and including the latest amplitude representative value A in measurement order. Process and
The vibration generated by the rotating machine is measured for a predetermined time at predetermined time intervals using the vibration detection sensor, and an amplitude representative value B that defines the amplitude distribution S is obtained from the amplitude distribution S, and the second And forming a new second data group by removing the oldest amplitude representative value A from the second data group and obtaining its distribution index DB;
When the distribution index DA of the second data group substantially matches the distribution index DB of the new second data group, the amplitude representative value B is set as the latest amplitude representative value A in the first data group. A step of deleting the amplitude representative value B when the distribution index DA of the second data group does not substantially match the distribution index DB of the new second data group. A method for diagnosing machine deterioration. 2. The method for diagnosing deterioration of a rotating machine according to claim 1, wherein the amplitude distributions R and S are obtained by dividing the vibration into a plurality of frequency regions and obtaining the vibrations for each of the frequency regions. . 3. The method for diagnosing deterioration of a rotating machine according to claim 2, wherein the average values of the amplitude distributions R and S are used as the amplitude representative values A and B, and the average values are used as the distribution indexes DA and DB. A method of diagnosing deterioration of a rotating machine, which employs a process capability index C _pi indicating a degree of variation of the rotating machine. 4. The method of diagnosing deterioration of a rotating machine according to claim 3, wherein a predetermined number of the representative amplitude values A are successively extracted from the first data group to form a plurality of third data groups. The representative value and the statistical index of the distribution of each third data group are sequentially obtained for each data group, the deterioration level is determined from the magnitude of each representative value, and the change of the statistical index of each third data group is determined. A method for diagnosing deterioration of a rotating machine, wherein the maintenance content of the rotating machine is determined from a speed. 5. The method for diagnosing deterioration of a rotating machine according to claim 4, wherein a central value M of each of the third data groups is adopted as the representative value, and an average value of the third data group is used as the statistical index. A method of diagnosing deterioration of a rotating machine, which employs a process capability index C _pik in consideration of the deviation of the rotation machine. 6. The method for diagnosing deterioration of a rotating machine according to claim 5, wherein an upper limit value and a lower limit value are set in advance for each of the median values M when calculating the process capability index C _pik. A method for diagnosing deterioration of rotating machinery. 7. The method for diagnosing deterioration of a rotating machine according to claim 6, wherein a frequency range to which the median M belongs and a rate of change of a process capability index C _pik corresponding to the median M are determined for each of the _medians M.
(1) Correspondence between each frequency region and the degradation type, (2) Correspondence between each degradation level included in each degradation type and the range of the median value, (3) Included in each degradation type set in advance Each maintenance action guideline within each deterioration level is compared with the corresponding relationship between the range of the rate of change of the process capability index C _pik ,
A method for diagnosing deterioration of a rotating machine, comprising: specifying a deterioration type and the deterioration level; and determining the specific maintenance contents.

Images