JP2018179986A - Abnormality diagnostic device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine the cause of detected abnormalities generated in a driving apparatus for a railway vehicle.SOLUTION: An abnormality diagnostic device 5 classifies results of octave-band analysis of vibration data detected by a vibration sensor 3 set in a railway vehicle, into at least three groups of data according to the low frequency band, the middle frequency band, and the high frequency band, and determines whether each group of data matches learning data 63 of the corresponding frequency band, and determines the presence of an abnormality in the frequency band. The device further determines the cause of an abnormality determined on the basis of the presence or absence of an abnormality in each frequency band.SELECTED DRAWING: Figure 4

Description

  The present invention relates to an abnormality diagnosis apparatus and the like using vibration data.

  Rail vehicles are regularly inspected to maintain safety. Further, for the purpose of promptly detecting an abnormality during operation and preventing an accident in advance, a technology for monitoring the state of various devices and parts mounted on a railway vehicle has been developed and put into practical use. As a method of state monitoring, a method of attaching various sensors such as a temperature sensor and a vibration sensor to each of devices and parts to be diagnosed is generally used (see, for example, Patent Document 1).

JP 2008-148466 A

  Among the equipment to be inspected, the equipment to be inspected particularly in a railway vehicle is a driving equipment. As an apparatus for driving a railway vehicle, in the case of an electric vehicle, a rotating machine such as a main motor and a reduction gear, and in the case of a pneumatic vehicle, a diesel engine, a transmission, a reduction gear, or a propulsion shaft is used. Since these drive devices are important devices that will prevent the safe and normal operation of the train if a failure occurs, it is important to detect the abnormality early and prevent the failure or damage in advance. is there. Further, there are various types of failures that may occur in the driving device, but when a suspicious abnormality occurs, it is desirable to reliably detect it without missing it before it reaches the failure.

  Furthermore, not only detecting the occurrence of an abnormality but also identifying the cause of the abnormality or narrowing down the cause of the abnormality is required to determine the cause of the abnormality. Even if the occurrence of an abnormality can be detected, if the cause of the abnormality is not known, it does not become a clue for a part requiring maintenance or repair, and it is difficult to appropriately cope with the generated abnormality. In addition, if it is known whether the event that is causing the abnormality is progressing / deteriorating, it is useful to be able to roughly know the time delay until the failure.

  The present invention has been made in view of the above circumstances, and has a first object to provide a technology capable of detecting an abnormality that has occurred in a driving device for a railway vehicle and determining the cause of the abnormality. Another object of the present invention is to provide a technique capable of determining whether an event that is causing an abnormality is progressing / deteriorating.

A first invention for solving the above-mentioned problems is
An analysis unit (for example, an octave band analysis unit 531 in FIG. 11) that performs octave band analysis on vibration data detected by a vibration sensor installed on a railway vehicle;
A classification unit (for example, a frequency band classification unit 535 in FIG. 11) that classifies the analysis result of the analysis unit into frequency band specific data of at least three frequency bands of low frequency band, middle frequency band, and high frequency band;
Abnormality determination means (for example, the abnormality determination unit in FIG. 11) that determines presence / absence of abnormality of the frequency band by determining whether each of the data classified by frequency band conforms to predetermined reference data of the corresponding frequency band 541),
Abnormality cause determination means (for example, an abnormality cause determination unit 547 in FIG. 11) for determining an abnormality cause determined based on the presence or absence of an abnormality in each of the frequency bands determined by the abnormality determination means;
An abnormality diagnosis device provided with

The second invention is
An analysis unit (for example, an octave band analysis unit 531 in FIG. 11) that performs octave band analysis on vibration data detected by a vibration sensor installed on a railway vehicle;
A classification unit (for example, a frequency band classification unit 535 in FIG. 11) that classifies the analysis result of the analysis unit into frequency band specific data of at least three frequency bands of low frequency band, middle frequency band, and high frequency band;
Abnormality determination means (for example, as shown in FIG. 11) which determines presence / absence of abnormality of the frequency band by determining whether each of the frequency band-specific data conforms to the predetermined reference data of the corresponding frequency band in a predetermined cycle. Abnormality determination unit 541),
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means (for example, abnormality degree history display control unit 549 shown in FIG. 11);
An abnormality diagnosis device provided with

Also, as another invention,
Computer,
Analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle,
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
An abnormality determination unit that determines presence or absence of an abnormality of the frequency band by determining whether each of the frequency band classified data conforms to predetermined reference data of the corresponding frequency band,
Abnormality cause determining means for determining an abnormality cause determined based on the presence or absence of an abnormality in each of the frequency bands determined by the abnormality determining means;
You may configure a program to function as

Also, as another invention,
Computer,
Analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle,
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
An abnormality determination unit that determines presence or absence of an abnormality of the frequency band by determining whether each of the data according to frequency band conforms to predetermined reference data of the corresponding frequency band at a predetermined cycle,
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means,
You may configure a program to function as

  When there is a mechanical abnormality in the railway vehicle, an abnormal vibration occurs due to it. And such abnormal vibration appears in the same frequency band if the cause of the abnormality is the same. According to the first and second inventions, etc., octave band analysis results of vibration data are classified and classified into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band. By determining whether each of the frequency band data conforms to the predetermined reference data of the corresponding frequency band, it is possible to determine the presence or absence of an abnormality in each frequency band. Then, according to the first aspect of the invention, the cause of the abnormality of the generated abnormality can be determined based on the presence or absence of the abnormality of each of the determined frequency bands. Further, according to the second aspect of the present invention, it is possible to display the ratio of the number of times that it is determined that there is an abnormality between a predetermined unit time or a predetermined traveling unit as a history. In this way, it can be determined whether or not the event that is causing the abnormality is progressing / deteriorating, and to what extent it is progressing / deteriorating.

Also, for the first invention described above, as in the second invention,
The abnormality determination means determines the presence or absence of each of the frequency band data in a predetermined cycle,
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means,
Further, the abnormality diagnosis apparatus as the third invention can be configured.

Also, as a fourth invention,
The classification unit classifies the upper limit threshold of the low frequency band as a frequency between 50 and 100 Hz.
The abnormality diagnosis apparatus of any of the first to third inventions may be configured.

  According to the fourth invention, the upper limit threshold of the low frequency band when classifying the octave band analysis result into the frequency band classified data can be a frequency between 50 and 100 Hz.

As a fifth invention,
The classification unit classifies data into three frequency bands of low frequency band, middle frequency band and high frequency band, and sets the lower threshold of the high frequency band to 1 kHz.
The abnormality diagnosis apparatus of any of the first to fourth inventions may be configured.

  According to the fifth invention, it is possible to set the lower limit threshold of the high frequency band to 1 kHz when classifying the octave band analysis result into the frequency band classified data.

Also, as a sixth invention,
The abnormality determination means has a means for calculating the degree of abnormality of the frequency band by calculating the degree of matching with the predetermined reference data of the corresponding frequency band for each of the data classified by frequency band,
The second history display control means (for example, the abnormality degree history display control unit 549 shown in FIG. 11) for performing control to display the abnormality degree historically is further provided.
The abnormality diagnosis apparatus of any of the first to fifth inventions may be configured.

  According to the sixth aspect of the invention, it is possible to calculate, as the degree of abnormality of the frequency band, the degree of conformity of the data of each frequency band with the corresponding predetermined reference data, and display the history of the degree of abnormality.

As a seventh invention,
The second history display control means identifies and displays the degree of abnormality in the display of the history.
You may comprise the abnormality-diagnosis apparatus of 6th invention.

  According to the seventh invention, it is possible to identify and display the degree of abnormality and display the history of the degree of abnormality.

The schematic diagram which shows the example of a whole structure of a state diagnostic system. The figure explaining the state diagnosis by abnormality diagnosis device. The figure explaining the octave band analysis of vibration data. The figure which shows the general flow of immediate diagnosis. The figure explaining the principle of abnormality degree calculation by the neighborhood method (NNDD). The figure which shows the abnormality cause discrimination | determination table used for discrimination | determination of abnormality cause. The figure which shows an example of the installation part classified abnormality degree log | history screen. The figure which shows an example of a run curve screen. The figure which shows the abnormal-generation rate in an abnormal wear test. The figure which shows an example of an abnormality incidence rate display screen. FIG. 2 is a block diagram showing an example of the functional configuration of an abnormality diagnosis device. The flowchart which shows the flow of immediate diagnosis processing. The flowchart which shows the flow of 1 run worth diagnostic processing.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. The present invention is not limited by the embodiments described below, and the mode to which the present invention can be applied is not limited to the following embodiments. Further, in the description of the drawings, the same portions are denoted by the same reference numerals.

  FIG. 1 is a schematic view showing an example of the entire configuration of a state diagnosis system 1 of the present embodiment. As shown in FIG. 1, the state diagnosis system 1 is configured to include a vibration sensor 3 and an abnormality diagnosis device 5, and is mounted on a railway vehicle 9 and used. The vibration sensor 3 is installed in the vicinity of the driving device of the railway vehicle 9 and detects the vibration generated by the operation of the driving device. In FIG. 1, the example installed in a trolley | bogie is shown. Then, the abnormality diagnosis device 5 diagnoses the state of the driving device based on the vibration data detected by the vibration sensor 3. For example, when the railway vehicle 9 is an electric vehicle, drive devices such as a main motor (motor), a transmission, and a drive device such as a gear device, bearings used for these, and the like are targets of state diagnosis. In the case of a diesel (diesel) vehicle, driving devices such as a diesel engine (engine), a transmission, a reduction gear, an accessory drive device, and peripheral components thereof are targets of state diagnosis. In addition, since the vibration sensor 3 can be installed for each of the main devices constituting the driving device, for example, in the case of a diesel car, an engine, a transmission, an accessory drive device, a radiator, etc. ) Can be installed.

〔principle〕
FIG. 2 is a diagram for explaining state diagnosis by the abnormality diagnosis device 5. The abnormality diagnosis device 5 processes the vibration data and determines whether the vibration data includes abnormal vibration as the presence or absence of abnormality. Then, the determination of the cause of the abnormality (the identification or narrowing down of the cause of the abnormality) using the determined presence or absence of abnormality is performed, and the state diagnosis of the driving device is performed. When a plurality of vibration sensors 3 are installed, vibration data is processed for each installation location, and the state diagnosis is performed.

  In the present embodiment, the abnormality diagnosis device 5 inputs vibration data from the vibration sensor 3 while the railway vehicle 9 is traveling, and performs octave band analysis (A1). Then, state diagnosis (instant diagnosis) is performed in real time using octave band analysis results for each predetermined unit period (for example, one second), and an immediate diagnosis result 707 is obtained for each unit period (A3).

  In addition, after completing one run, state diagnosis (diagnosis for one run) is performed based on the immediate diagnosis result for one run, and one run diagnosis 717 is obtained for each run (A5). Here, one run (one run) is a run over a period of a certain length, and can be, for example, a run from a departure station to an end station, or a run on a day. If it is defined from the viewpoint of vehicle operation, one traveling route or work may be run once, or operation for one train to which the vehicle is applied may be run once.

  Further, control is performed to receive a predetermined display instruction operation and display an immediate diagnosis result and a 1-run diagnosis result on the display unit 55 (see FIG. 11) and the like of the abnormality diagnosis device 5 (A7).

(1) Octave Band Analysis FIG. 3 is a diagram for explaining octave band analysis of vibration data. In the octave band analysis, predetermined filtering processing is continuously performed on the vibration data D11, and the octave band analysis result is recorded for each unit period Δt. Thereby, the magnitude (vibration effective value) of the vibration for each frequency band (octave band) is obtained as the octave band analysis result D13 in each unit period Δt. In the present embodiment, the average value (average speed) of the traveling speeds of the railway vehicle 9 in the unit period Δt in which the octave band analysis is performed is calculated as the traveling speed corresponding to the octave band analysis result D13. In addition, the number of rotations of the power source (motor or engine) and the operation mode in the unit period Δt are acquired, and recorded along with the calculated traveling speed in association with the octave band analysis result D13. The operation mode is a combination of the driving operation of power running, coasting and brake, and the number of notches thereof. For example, power running 5 notch · · · power running 1 notch, coasting, brake 1 notch · · · brake 5 notch Is set as

  Although octave band analysis is performed in real time while traveling as in the present embodiment from the viewpoint of data amount reduction, a method of recording only octave band analysis results for each unit period Δt without recording vibration data It is suitable to adopt. However, it is also possible to record vibration data related to one run and perform octave band analysis for each unit period Δt later.

(2) Immediate Diagnosis FIG. 4 is a diagram showing a rough flow of the immediate diagnosis. In the immediate diagnosis, first, the octave band analysis result of the unit period (target unit period) to be diagnosed is classified into frequency band data of three frequency bands of low frequency band, middle frequency band and high frequency band (step A31) . Specifically, as shown in FIG. 3, octave bands are divided into low frequency bands, middle frequency bands and high frequency bands, and vibration effective values of octave bands belonging to each frequency band correspond to frequency bands. Classified as frequency band data. Then, a feature vector having the classified vibration effective value as an element is obtained as a feature vector according to the corresponding frequency band. Alternatively, principal component analysis and whitening of the frequency band data may be performed to determine the principal component, and the determined principal component may be used as a feature vector. The number of main components can be determined according to the contribution rate. The specific range of each frequency band will be described later with reference to FIG. 6, but it may be classified into four or more frequency bands more finely than that.

  Subsequently, as shown in FIG. 4, learning data prepared in advance for each frequency band as predetermined reference data for each classified frequency band data of each of the classified low frequency band, middle frequency band and high frequency band. It is determined whether or not there is an abnormality in each frequency band by determining whether or not it matches the learning data 63 of the corresponding frequency band among 63 (631, 633, 635) (step A33).

  First, learning data [low frequency band (normal)] 631 is an octave band analysis result for normal vibration data collected in advance, and is a set of feature vectors based on frequency band data of the low frequency band thereof. is there. Similarly, the learning data [middle frequency band (normal)] 633 is an octave band analysis result for vibration data at normal time, and is a set of feature vectors based on data in the middle frequency band. The learning data [high frequency band (normal)] 635 is an octave band analysis result for normal vibration data, and is a set of feature vectors based on high frequency band frequency band data.

  Further, for determination, for example, NNDD (Nearest Neighbor Data Description), which is a type of a proximity method, or a one class support vector machine can be used. FIG. 5 is a diagram for explaining the principle of abnormality degree calculation by the neighborhood method (NNDD). As shown in FIG. 5, in NNDD, frequency band data (more specifically, a feature vector obtained from the frequency band data) is considered as one point on a multi-dimensional space. Then, in the multi-dimensional space, from among the set G of learning data (feature vectors based on vibration data at the normal time), learning data A closest to the test data X (feature vector to be determined) representing the vibration data is searched A ratio obtained by dividing the distance d1 between the test data X and the learning data A by the reference distance is obtained to determine whether the test data X is normal or abnormal. When the test data X is abnormal, it is considered that the distance between the learning data A and the test data X becomes large, and the obtained ratio represents the degree of abnormality (abnormality degree).

In actual processing, each of the feature vectors obtained for each frequency in the preceding step A31 is sequentially processed. Then, not only the learning data of the nearest neighbor but also the learning data close to the k-th (k = 1 to k _NN ) is used, and the feature vector of the judgment target is the corresponding learning data 63, that is, normal time related to the corresponding frequency band It is determined whether the feature vector of each frequency band is normal or abnormal, respectively, in comparison with the feature vector based on the vibration data of.

ここで、ＮＮ_k（ｘ）は、データｘにｋ番目に近い学習データを表している。 Specifically, the abnormality degree g of the feature vector x to be determined is obtained according to the abnormality degree calculation function g (x) shown in the following equation (1). In equation (1), d _k represents a reference distance. The reference distance d _k is determined as follows. That is, for each feature vector of the learning data 63, the distance to the k-th closest feature vector is calculated. Then, by arranging the calculated distance in ascending order to identify the distance between the feature vectors located 99% of the ranking, which is a reference distance d _k according to the feature vector is close to the k-th.

Here, NN _k (x) represents learning data that is k-th closest to data x.

  Then, if the abnormality degree calculated by the abnormality degree calculation function g (x) in equation (1) is a positive value, the feature vector to be determined is determined to be abnormal, and if the abnormality degree is a negative value, the feature vector to be determined Is determined to be normal. When it is determined that the feature vector is abnormal, the vibration data of the frequency band is determined as "abnormal" as it includes abnormal vibration. On the other hand, when it is determined that the feature vector is normal, the vibration data of the frequency band is determined to be "abnormal" since it does not include abnormal vibration.

  Next, the case of using a one-class support vector machine will be briefly described. One-class support vector machine is a method that applies a support vector machine, which is a learning model for classifying feature vectors into two classes (groups). The support vector machine determines a hyperplane for classifying learning data whose class is defined such that the distance (margin) between two classes of data is maximum, and the feature vector of the determination target is either class according to this hyperplane Classified into Then, the one-class support vector machine determines a hyperplane that classifies the class of learning data and the other using only one class of normal data as learning data, and determines whether the feature vector to be determined is normal according to the determined hyperplane Classify as either abnormal.

In the one-class support vector machine, the discrimination function f (x) used to determine whether the feature vector x to be determined is normal or abnormal is determined by the following expression (2). When the discrimination function f (x) is a positive value, the feature vector x to be determined is not classified into the class of normal data as learning data, that is, it is abnormal. On the other hand, when the discriminant function f (x) is a negative value, the feature vector x to be determined is classified into a class of normal data which is learning data, that is, it is normal. "L" is the number of feature vectors that make up the training data.

In the equation (2), the function k (x, y) is a kernel function, which is determined by the following equation (3) using a Gaussian kernel. In equation (3), the parameter σ is determined, for example, as the maximum value of the distance between feature vectors.

Also, in equation (2), the vector x _i is a support vector. The support vector is a feature vector whose corresponding coefficient α is not “0” among the feature vectors x of normal data. Coefficient alpha _i of support vector x _i is calculated as a solution of the minimization problem of equation (4). In equation (4), the parameter は is set to, for example, “0.1”.

Further, the parameter 式 of the equation (2) is given by the following equation (5) when the coefficient α _i of the support vector x _i is neither the upper limit nor the lower limit.

  Also in the case of using such a one-class support vector machine method, each of the feature vectors obtained for each frequency band in step A31 in the previous stage is sequentially determined. Then, the feature vector to be determined is compared with the corresponding learning data 63 to determine whether the feature vector of each frequency band is normal or abnormal.

Specifically, an abnormality degree calculation function g (x) shown in the following equation (6) is determined using the discriminant function f (x), and according to this abnormality degree calculation function g (x), Find the degree of abnormality g.

  The abnormality degree calculation function g (x) in equation (6) is a value obtained by normalizing the value obtained by the discriminant function f (x) with the parameter ρ. It can be said that the evaluation value is normalized by the one-class support vector machine. Therefore, if the abnormality degree is a positive value, the feature vector to be determined is determined to be abnormal, and if the abnormality degree is a negative value, the feature vector to be determined is determined to be normal. When it is determined that the feature vector is abnormal, the vibration data of the frequency band is determined as "abnormal" as it includes abnormal vibration. On the other hand, when it is determined that the feature vector is normal, the vibration data of the frequency band is determined to be "abnormal" since it does not include abnormal vibration.

  Subsequently, the cause of the abnormality is determined based on the presence or absence of the abnormality in each of the determined frequency bands (step A35). FIG. 6 is a diagram showing an abnormality cause determination table used to determine an abnormality cause. The determination of the cause of abnormality according to the present embodiment does not accurately identify the part in which the abnormality has occurred, but presumesly identifies or narrows down the event that is the cause of the abnormality. If the event causing the abnormality can be identified, it is possible to narrow down the parts that may have an abnormality, the inspection target becomes clear, and the final identification, maintenance, and repair of the abnormal parts are made. Can be used to improve work efficiency and schedule development.

  Specifically, as shown in FIG. 6, when it is determined that there is an abnormality in any of the low frequency band, the middle frequency band, and the high frequency band, according to the combination of the presence or absence of the abnormality in each frequency band , Unevenness or misalignment (rotational misalignment) of the rotating machine that constitutes the drive device, looseness due to tightening failure of the fastening member between the parts that constitute the drive device, etc. Abnormal wear due to etc., and motor abnormalities are determined from the corresponding abnormal causes.

  Here, as shown in FIG. 6, imbalance or misalignment is one of the causes of abnormality that is determined when it is determined that there is an abnormality in the low frequency band. Anomalous vibration due to unbalance or misalignment fluctuates on the low frequency side according to the rotational speed of the rotary machine. For example, in the case of an electric car, the maximum number of revolutions of the motor is about 6,000 revolutions / min (maximum revolution frequency 100 Hz), and in the case of a diesel car, the maximum number of revolutions of the engine is 2,000 revolutions / min (maximum revolution frequency 33. 3 Hz). In the unbalance, vibrations corresponding to this rotational frequency occur. Further, when the power transmission shaft is a cardan shaft using a universal joint as in a diesel car and there is misalignment, vibration of twice the rotational frequency occurs. Therefore, in the classification step A31 of the frequency-classified data of the preceding stage, the low frequency band is classified with the upper threshold of the low frequency band as a frequency between 50 Hz and 100 Hz. In the present embodiment, for example, the upper limit threshold is set to 100 Hz. The lower limit threshold of the high frequency band is 1 kHz, and the middle frequency band and the high frequency band are classified.

  The result of the immediate diagnosis is recorded as the immediate diagnosis result 707 (see FIG. 2). In the immediate diagnosis result 707, the presence or absence of abnormality of each determined frequency band, the degree of abnormality calculated in the determination, and the determined cause of abnormality are set.

  According to the above-described immediate diagnosis, it is possible to determine the presence or absence of an abnormality in each of the low frequency band, the medium frequency band, and the high frequency band, and to determine the cause of the abnormality. According to this, it is possible to narrow down and inspect a part which may have an abnormality from among the corresponding parts constituting the drive device, based on the determined cause of abnormality. Specifically, as shown in FIG. 6, when the determined cause of abnormality includes "unbalance or misalignment", an appearance inspection or a dimensional inspection is performed. In addition, when it includes "looseness", a hammering sound test is performed, when it includes "abnormal wear", wear powder analysis is performed, and when it includes "motor abnormality", a motor inspection is performed. And since maintenance and repair as needed can be performed based on the result of these inspections, it becomes possible to cope with the generated abnormality appropriately and efficiently.

  The immediate diagnosis described here is performed only for the octave band analysis result based on the vibration data when the drive device is in the steady state, and the octave band analysis when the operation of the drive device is in the transition state The result may not be performed.

  For example, the octave band analysis results before and after in time are compared to determine whether or not the steady state is obtained. Specifically, the difference between a certain octave band analysis result and the octave band analysis result immediately before in time is determined, and if the difference is less than a predetermined threshold, the steady state is determined, and if it exceeds the threshold, the transient state is determined. Do. Here, the difference between octave band analysis results is the difference between vibration effective values for each octave band. For example, when the square root of the sum of squares of the difference of vibration effective values of each octave band exceeds a threshold, it is determined that the transient state is present.

  Further, learning data for each frequency band of low frequency band, middle frequency band and high frequency band may be prepared for each operation mode. Then, with reference to learning data corresponding to the operation mode in the target unit period, the presence or absence of abnormality in each frequency band may be determined.

(3) Diagnosis for one run In the diagnosis for one run, at an appropriate timing after completing one run, the degree of abnormality for each frequency band calculated in the immediate diagnosis during the run is averaged, and one run is calculated. The average abnormality degree is calculated for each frequency band. Then, the presence or absence of an abnormality in each frequency band is determined according to the calculated average abnormality degree, and the cause of the abnormality is determined based on the presence or absence of an abnormality in each determined frequency band. The determination of the presence or absence of the abnormality and the determination of the cause of the abnormality can be performed in the same manner as the immediate diagnosis. That is, if the average abnormality degree is a positive value, it is determined as "abnormal", and if it is a negative value, it is determined as "no abnormality". Then, using the abnormality cause determination table shown in FIG. 6, the cause of the abnormality is determined based on the presence or absence of an abnormality in each frequency band.

  The result of the one running diagnosis is recorded as the one running diagnosis 717 (see FIG. 2). In the one running diagnostic result 717, the presence or absence of an abnormality in each determined frequency band, the average abnormality degree calculated in the determination, and the determined cause of the abnormality are set.

(4) Display Control of Diagnosis Results The abnormality diagnosis apparatus 5 performs control of displaying the diagnosis results of the above-described immediate diagnosis and one running diagnosis on the display unit 55 or the like, and presents the display to the user. Further, control is performed to display the history of the degree of abnormality calculated in the immediate diagnosis and the one running diagnosis for each installation location of the vibration sensor 3.

  FIG. 7 is a view showing an example of the installation position classified abnormality degree history screen W1. As shown in FIG. 7, the history of the degree of abnormality calculated for the installation location of the vibration sensor 3 to be displayed is displayed, for example, every day, every week, or every month on the installation location classified abnormality level history screen W1. Displayed in tabular form as a unit. FIG. 7 shows an example of a screen when each day is selected as the display unit.

  When displaying the abnormality degree history screen W1 classified by installation location, for example, when displaying the abnormality degree history every day, first, based on the immediate diagnosis result according to the installation location of the vibration sensor 3 to be displayed The average value of the abnormality degree in the unit period is calculated on a daily basis for each of the low frequency band, middle frequency band, and high frequency band, and the daily average abnormality degree of each date for each frequency band is calculated. In addition, when it diagnoses by making 1 driving | running | working into driving | running | working of 1st, the average abnormality degree calculated by 1 driving part diagnosis which concerns on the said installation location can be used as a daily average abnormality degree. In addition, when the display unit for each week or month is selected, the average value for each week or month of the immediate diagnosis result relating to the installation location may be calculated.

  Subsequently, the calculated daily average abnormality degree is subjected to threshold determination, and the degree of abnormality is determined in four stages of, for example, “no abnormality”, “abnormality level low”, “medium abnormality level”, and “high abnormality level”. Specifically, if the daily average abnormality degree is a negative value, "no abnormality" is set. On the other hand, when the daily average abnormality degree is a positive value, for example, a threshold that divides the range of possible values of the abnormality degree into three levels is defined in advance. However, the number of stages is not limited to three, and may be two or four or more. Then, the daily average abnormality degree is thresholded using the threshold value, and the degree of abnormality (high in abnormality degree level, low in abnormality degree level, or high in abnormality degree level) is determined.

  When the degree of abnormality of each day is determined, the daily average abnormality degree of each frequency band is displayed in the display column of the corresponding date on the installation place-classified abnormality degree history screen W1. At that time, by setting the background color of the display column as the background color according to the degree of abnormality determined for the daily average abnormality degree of the date, the level of the abnormality degree is identified and displayed.

  According to the display of the installation part-by-place abnormality degree history screen W1, the abnormality degree history such as each day in each frequency band can be displayed, and the level of each abnormality can be identified and displayed in the display of the abnormality degree history. According to this, the user can easily grasp the change in height together with the degree of abnormality such as daily. In addition, if the degree of abnormality increases gradually, it is considered that the event of the abnormality cause is progressing / deteriorating, so the time delay from the display of the abnormality degree history to the failure is roughly To know. Therefore, it is possible to reliably cope with the abnormality that has occurred, such as adjusting the timing of maintenance and repair of the railway vehicle 9, before the failure.

  In addition, the display mode of the abnormality degree history is not limited to the mode in which the average value of each day, week, month, etc. shown in FIG. 7 is displayed in tabular form, for example, the average value of each day etc. , And may be displayed as a graph.

  In addition, the history of the degree of abnormality calculated for each unit period in the immediate diagnosis can be displayed for each installation location. FIG. 8 is a view showing an example of the run curve screen W2 including the display of the abnormality degree history for each unit period in each frequency band. As shown in FIG. 8, the run curve screen W2 is configured by arranging the abnormality degree display unit W21, the notch display unit W22, the rotation speed display unit W23, and the traveling speed display unit W24 vertically, The abnormality degree history of each frequency band for traveling is displayed on the same traveling position coordinate axis or the same time axis together with the notch, the number of revolutions, and the traveling speed. Further, the run curve screen W2 has a slider S2 which is slidable along the left and right traveling position coordinate axis directions, and the degree of abnormality, notch, and the like of each frequency band at the timing of traveling the traveling position indicated on the slider S2. Display each value of rotation speed and traveling speed. In FIG. 8, the display on the slider S2 is omitted. Further, FIG. 8 exemplifies the run curve screen W2 when the railway vehicle 9 is a pneumatic vehicle, and the engine speed (solid line) indicating the rotational speed of the engine and the output of the transmission are displayed in the engine speed display portion W23. An axis rotation number (broken line) indicating the rotation number of the axis is displayed. In the case of an electric car, the rotational speed display unit W23 displays the rotational speed of the motor.

  Further, the display of the abnormality degree history may be as follows. That is, for each frequency band, the number of times it is determined that there is an abnormality in immediate diagnosis during a predetermined unit time such as one day or one week, or a predetermined travel unit such as one travel or one vehicle operation. Display percentages historically.

  FIG. 9 is a diagram historically showing the abnormal occurrence rate in the abnormal wear test of the engine carried out for the verification of the abnormal vibration detection. This abnormal wear test is a test that simulates a process in which abnormal wear progresses / deteriorates by gradually mixing ceramic powder (SiC: silicon carbide) into engine lubricating oil. The stepwise mixing of the SiC powder means that the concentration of the SiC powder increases with the passage of time, that is, the passage of time during which abnormal wear progresses / deteriorates. In addition, when the concentration of the SiC powder reached 0.5%, the engine was disassembled and examined, and it was confirmed that the continuous use of the engine was not permitted and it was in a broken state.

  The abnormality occurrence rate is a ratio of the number of times of performing the same diagnosis as the immediate diagnosis of the above-described embodiment and determining that there is an abnormality (= the abnormality degree is a positive value). The numerical magnitude of the degree of abnormality is not critical. The ratio of the number of times that it is determined that there is an abnormality in the immediate diagnosis at each stage where the concentration of the SiC powder is the same is the abnormality occurrence rate. The anomaly incidence rate was calculated for each of three frequency bands: low frequency band, middle frequency band and high frequency band.

  As a result, as shown in FIG. 9, it was found that the abnormality occurrence rate tends to increase as the concentration of the SiC powder becomes higher, in all of the low frequency band, the medium frequency band and the high frequency band. . For example, in a rotating machine or the like, even if it is determined that there is an abnormality at a certain moment, abnormal vibration may be settled at the next moment and it may be determined that there is no abnormality. However, when the concentration of the SiC powder increases, that is, as the abnormal wear progresses / deteriorates, the frequency of occurrence of the abnormal vibration increases and finally continues without occurrence of the abnormal vibration (= abnormal occurrence rate 100% It turns out that). Although the above-mentioned average abnormality degree and daily average abnormality degree are useful as an index showing that abnormal wear progresses / deteriorates, it can be said that this abnormal incidence rate is also useful.

  In addition, according to the frequency band, the frequency band where the abnormality occurrence rate tends to increase early as it is determined that there is an abnormality, or the frequency band where the abnormality incidence rate tends to increase at the end of approaching failure It was also found that there is a difference in the upward trend. However, depending on the installation location of the vibration sensor 3 or the device to be diagnosed, a difference may occur in the rising tendency.

  Although the display method of the abnormality occurrence rate can be considered as appropriate, it is important that it is a historical display. For example, since the horizontal axis in FIG. 9 can be read as the elapsed time as it is, it can be displayed as in the abnormality occurrence rate display screen W3 shown in FIG. The horizontal axis can be a time axis with a predetermined unit time, such as one day or one week, or a predetermined travel unit, such as one travel or one vehicle operation, as one scale. The abnormality occurrence rate display screen W3 is an example of the abnormality degree history display.

  By displaying the abnormality occurrence rate of each frequency band historically, if the abnormality occurrence rate is getting higher gradually, it is considered that the event of the abnormality cause is progressing / deteriorating, leading to the failure You can get a rough idea of the temporal grace of Therefore, it is possible to reliably cope with the abnormality that has occurred, such as adjusting the timing of maintenance and repair of the railway vehicle 9, before the failure.

  Although the display mode of the abnormality occurrence rate is a bar graph in FIG. 10, it may be displayed as a line graph. Further, when there are a plurality of installation locations of the vibration sensor 3 and a plurality of devices to be diagnosed, it is preferable to switch and display. In addition, a threshold may be provided for the abnormality occurrence rate, and when the abnormality occurrence rate becomes equal to or more than the threshold, a notification or a warning may be output for alerting or warning. In this case, it is preferable to notify of which frequency band the abnormality occurrence rate satisfies the threshold condition.

  In addition, when calculating the rate of occurrence of abnormality, it is not intended to target all data of a predetermined unit time or a predetermined unit travel but only data when the device is operating in a predetermined operation mode. An abnormality occurrence rate may be calculated. Depending on the type of abnormality, abnormal vibration may occur only when the device is operating in a predetermined operation mode. In such a case, it is preferable to calculate the abnormality occurrence rate only for data when the device is operating in the predetermined operation mode.

[Functional configuration]
FIG. 11 is a block diagram showing an example of a functional configuration of the abnormality diagnosis device 5. As shown in FIG. 11, the abnormality diagnosis apparatus 5 includes an operation input unit 51, a processing unit 53, a display unit 55, a sound output unit 57, a communication unit 59, and a storage unit 60, and is a kind of computer Configured as

  The operation input unit 51 is for inputting necessary various operations, and can be realized by an input device such as a keyboard, a mouse, a touch panel, and various switches.

  The processing unit 53 is realized by an arithmetic device such as a CPU, for example, and performs data input / output control with the respective units of the apparatus including the operation input unit 51 and the storage unit 60. Then, the processing unit 53 performs various arithmetic processing based on the program and data stored in the storage unit 60, the operation input signal from the operation input unit 51, the data received from the outside through the communication unit 59, etc. An instruction and data transfer to each part constituting the abnormality diagnosis device 5 are performed to control the abnormality diagnosis device 5 in an integrated manner.

  Further, the processing unit 53 includes an octave band analysis unit 531, a frequency band classification unit 535, an abnormality determination unit 541, an abnormality cause determination unit 547, and an abnormality degree history display control unit 549.

  The octave band analysis unit 531 performs octave band analysis by performing predetermined filter processing on vibration data detected by the vibration sensor 3, and calculates a vibration effective value for each octave band for each predetermined unit period.

  The frequency band classification unit 535 uses the octave band analysis result of the octave band analysis unit 531 for each unit period in the implementation of the immediate diagnosis, and the data of each frequency band of low frequency band, middle frequency band and high frequency band. Classified into

  The abnormality determination unit 541 includes an immediate diagnosis determination unit 543 and a one travel diagnosis determination unit 545. The immediate diagnosis determination unit 543 determines, for each frequency band, whether the vibration data in the target unit period is normal or abnormal. Specifically, each of the frequency bands is determined whether or not the data classified by frequency band classified by the immediate diagnosis time classification unit 537 conforms to the learning data 63 (631, 633, 635) of the corresponding frequency band. Determine the presence or absence of an abnormality in Further, the one running diagnosis determination unit 545 averages the degree of abnormality for each frequency band calculated in the immediate diagnosis during one running, and each frequency according to the sign of the sign of the average degree of abnormality for the one running Determine if there is an abnormality in the band.

  The abnormality cause determination unit 547 determines the cause of the abnormality according to the abnormality cause determination table 65 based on the presence or absence of the abnormality in each of the frequency bands determined by the immediate diagnosis time determination unit 543 when performing the immediate diagnosis. Further, when performing the one running diagnosis, the cause of the abnormality is determined according to the abnormality cause determination table 65 based on the presence or absence of the abnormality in each of the frequency bands determined by the one running diagnosis determining unit 545.

  The abnormality degree history display control unit 549 performs control to historically display the abnormality degree by using the abnormality degree of each frequency band calculated by the immediate diagnosis in accordance with the display unit such as every day. At this time, the level of abnormality is determined, and the level of the determined abnormality is displayed in the display of the history of abnormality.

  Further, as described with reference to FIG. 10, the abnormality degree history display control unit 549 sets, for each frequency band, a predetermined unit time such as one day or one week, which is temporally longer than the cycle of the immediate diagnosis, or The control may be performed to historically display the ratio of the number of times it is determined that there is an abnormality in the immediate diagnosis during a predetermined traveling unit such as one traveling or one vehicle operation.

  The display unit 55 performs various displays based on the display signal from the processing unit 53. For example, it can be realized by a liquid crystal display or an organic EL display. The sound output unit 57 performs various sound outputs based on the sound signal from the processing unit 53. For example, it can be realized by a speaker or the like.

  The communication unit 59 performs data communication with an external device. For example, it can be realized by a wireless communication device, a modem, a TA (terminal adapter), a jack of a communication cable for wired communication, a control circuit, or the like.

  The storage unit 60 stores in advance a program for operating the abnormality diagnosis device 5 to realize various functions of the abnormality diagnosis device 5, data used during the execution of the program, or the like. It is temporarily stored each time. For example, it can be realized by an IC memory such as a RAM or a ROM, a memory card or a USB memory, a magnetic disk such as a hard disk, or an optical disk such as a CD-ROM or a DVD. In the storage unit 60, a diagnosis program 61, learning data 63, an abnormality cause determination table 65, data for one run 66, and diagnosis result data 69 are stored.

  The diagnostic program 61 is a program for causing the processing unit 53 to function as an octave band analysis unit 531, a frequency band classification unit 535, an abnormality determination unit 541, an abnormality cause determination unit 547, and an abnormality degree history display control unit 549.

  The learning data 63 includes learning data [low frequency band (normal)] 631, learning data [medium frequency band (normal)] 633, and learning data [high frequency band (normal)] 635.

  The abnormality cause determination table 65 is a data table which defines an abnormality cause corresponding to a combination of presence / absence of abnormality in each frequency band of low frequency band, middle frequency band and high frequency band (see FIG. 6).

  In one run data 66, analysis result data 68 generated by the octave band analysis unit 531 for each unit period during one run is accumulated for each run. In the one running data 66, the running date 661, the installation location 662 of the vibration sensor 3, and the one running data number 663 identifying the one running data 66 are set together. Further, each of the analysis result data 68 stored includes a unit period 681, an octave band analysis result 682, a traveling speed 683, a number of revolutions 685, and an operation mode 686. Note that when octave band analysis is not performed during traveling and only vibration data is recorded, vibration data in the corresponding unit period 681 is recorded instead of the octave band analysis result 682.

  The diagnosis result data 69 associates the immediate diagnosis result 707 (see FIG. 2) for each unit period obtained during one traveling with the traveling date 701, the installation location 703 of the vibration sensor 3 and the data number 705 for one traveling Immediate result group 70 accumulated in chronological order, and one run diagnostic result 717 (see FIG. 2) for each run, run date 711 and installation location 713 of vibration sensor 3, and one run data number 715 The one run result group 71 stored in association with one another is stored.

[Flow of processing]
The abnormality diagnosis device 5 executes a process (instant diagnosis process) relating to an immediate diagnosis for each unit period, and executes a process (one run portion diagnosis process) relating to a one run diagnosis after finishing one run Do. In addition, a predetermined display instruction operation is received, and a process (diagnosis result display control process) related to display control of a diagnosis result including display control of an abnormality degree history is executed. These processes are realized by the processing unit 53 reading out and executing the diagnostic program 61.

(1) Immediate Diagnosis Process FIG. 12 is a flowchart showing the flow of the immediate diagnosis process. In the instant diagnosis process, when the predetermined diagnosis start timing comes (step S1: YES), the octave band analysis unit 531 starts octave band analysis using vibration data from the vibration sensor 3 (step S3). By this processing, an octave band analysis result 682 is obtained for each unit period, and analysis result data 68 associated with the traveling speed 683, the number of revolutions 685, and the operation mode 686 in the unit period 681 is generated. It accumulates in one run data 66 related to the run.

  Here, the diagnosis start timing can be, for example, a timing at which a predetermined time (for example, 5 seconds) has elapsed after the departure from the stopping station. However, the present invention is not limited to this, and the timing may be such that the driving operation satisfies a predetermined driving operation condition, such as when the notch (the powering notch and / or the brake notch) becomes a predetermined notch or when traveling in an oval. Alternatively, a calculation device for separately calculating the traveling position may be installed on the vehicle, and the timing at which the vehicle passes a position satisfying a predetermined position condition (for example, a predetermined kilometer position) may be set as the diagnosis start timing.

  Once the octave band analysis is started, the processing of the loop L is performed for each unit period of the octave band analysis (steps S5 to S17). That is, in the loop L, first, the instantaneous diagnosis time classification unit 537 classifies the octave band analysis result 682 of the target unit period into frequency band data of each frequency band of low frequency band, middle frequency band, and high frequency band. (Step S9). Then, a feature vector of each frequency band representing each of the frequency band data, for example, a feature vector having a vibration effective value of an octave band classified as the frequency band data as a feature vector of the corresponding frequency band Obtained (step S11).

  Subsequently, the immediate diagnosis determination unit 543 determines whether or not each feature vector of each frequency band conforms to the learning data 63 of the corresponding frequency band, and determines the presence or absence of an abnormality in each frequency band (Step S13). Specifically, the abnormality degree for each frequency band is calculated according to the above equation (1) or (6), and the presence or absence of an abnormality in each frequency band is determined according to the sign of the sign.

  Then, according to the combination of presence / absence of abnormality of each frequency band determined in step S13, the abnormality cause determination unit 547 determines an abnormality cause corresponding to the combination with reference to the abnormality cause determination table 65 (step S15).

(2) One Traveling Part Diagnosis Process FIG. 13 is a flowchart showing a flow of the one traveling part diagnosis process. In the one running diagnostic process, first, the one running diagnostic time determination unit 545 averages, for each frequency band, the degree of abnormality of each frequency band set in the immediate diagnostic result 707 for one running target to be diagnosed. The average abnormality degree of each frequency band is calculated (step S21). Then, the presence or absence of an abnormality in each frequency band is determined according to the positive / negative of the calculated sign of the average abnormality degree (step S23). Thereafter, according to the combination of presence / absence of abnormality of each frequency band determined in step S23, the abnormality cause determination unit 547 determines an abnormality cause corresponding to the combination with reference to the abnormality cause determination table 65 (step S25).

(3) Diagnosis result display control processing In the diagnosis result display control processing, control is performed such that the abnormality degree history display control unit 549 displays the history of the abnormality degree for each frequency band calculated in the immediate diagnosis processing for each installation location of the vibration sensor 3 Do. Specifically, the degree of abnormality for each frequency band in each unit period calculated in step S13 of FIG. 12 is averaged, for example, for each day to calculate the average degree of daily abnormality, and the calculated average degree of abnormality for day is arranged and displayed in date order In this way, display and control the daily abnormal degree history. Further, the abnormality degree history display control unit 549 makes a threshold determination of the calculated daily average abnormality degree, and determines the degree of the abnormality degree. Then, in the display of the abnormality degree history, the level of the determined abnormality degree is identified and displayed.

  In addition, in the diagnosis result display control process, control for displaying the cause of the abnormality determined in step S15 of FIG. 12 in the immediate diagnosis process together with the presence or absence of the abnormality determined in step S13 is displayed on the display unit 55. Do. Further, control is performed to display on the display unit 55 the cause of abnormality determined in step S25 of FIG. 1 in the one running diagnostic process together with the presence or absence of the abnormality determined in step S23. At this time, a sound output unit 57 may output a predetermined notification sound.

  As described above, according to the present embodiment, it is possible to detect an abnormality that has occurred in a railway vehicle and to determine the cause of the abnormality.

  In the above embodiment, although an example in which the vibration sensor 3 is installed on one carriage in FIG. 1 is shown, the vibration sensor 3 may be installed on a drive equipment basis, for example, installed on all carriages. In that case, the device configuration of the abnormality diagnosis device 5 is not limited to the single configuration shown in FIG. 1 and may be realized by a plurality of computers sharing functions. For example, while installing a terminal device for performing processing related to octave band analysis for each vibration sensor, based on the result of octave band analysis in each terminal device, immediate diagnosis or diagnosis for one run for each carriage on which the vibration sensor 3 is installed Alternatively, a central device that performs time series diagnosis may be separately installed. In this case, the central unit is disposed on the ground such as a garage, and when the railway vehicle 9 is located in a garage or the like in which the central unit is disposed, it communicates with a terminal device installed on the railway vehicle 9 , Octave band analysis results can be obtained. In addition, maintenance personnel may manually move the data of the octave band analysis result from the terminal device to the central device via a storage medium such as a memory card.

  In addition, when only recording of vibration data is performed during traveling, the abnormality diagnosis device 5 mounted on the railway vehicle 9 performs state diagnosis equivalent to octave band analysis or immediate diagnosis after the operation such as night time, for 1 run Processing relating to diagnosis and time-series diagnosis may be performed, or a configuration may be employed in which an abnormality diagnosis device installed on the ground (outside the vehicle) processes the recorded vibration data and performs processing relating to octave band analysis and various diagnoses.

DESCRIPTION OF SYMBOLS 1 ... Condition diagnostic system 3 ... Vibration sensor 5 ... Abnormality diagnosis device 9 ... Rail vehicle 51 ... Operation input part 53 ... Processing part 531 ... Octave band analysis part 535 ... Frequency band classification part 541 ... Abnormality judgment part 543 ... Determination at the time of immediate diagnosis Section 545 ... 1 run time diagnosis section 547 ... abnormality cause discrimination section 549 ... abnormality degree history display control section 55 ... display section 57 ... sound output section 59 ... communication section 60 ... storage section 61 ... diagnostic program 63 (631, 633 , 635) ... Learning data 65 ... Abnormal cause determination table 66 ... 1 run data 68 ... Analysis result data 69 ... Diagnosis result data 707 ... Immediate diagnosis result 717 ... 1 run diagnosis result

Claims (9)

An analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle;
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
Abnormality determination means for determining presence / absence of abnormality of the frequency band by determining whether each of the frequency band classified data conforms to predetermined reference data of the corresponding frequency band;
Abnormality cause determination means for determining an abnormality cause determined based on the presence or absence of an abnormality in each of the frequency bands determined by the abnormality determination means;
An abnormality diagnostic device equipped with An analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle;
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
Abnormality determination means for determining presence or absence of an abnormality of the frequency band by determining whether each of the frequency band data conforms to the predetermined reference data of the corresponding frequency band in a predetermined cycle;
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means,
An abnormality diagnostic device equipped with The abnormality determination means determines the presence or absence of each of the frequency band data in a predetermined cycle,
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means,
The abnormality diagnosis device according to claim 1, further comprising: The classification unit classifies the upper limit threshold of the low frequency band as a frequency between 50 and 100 Hz.
The abnormality diagnosis device according to any one of claims 1 to 3. The classification unit classifies data into three frequency bands of low frequency band, middle frequency band and high frequency band, and sets the lower threshold of the high frequency band to 1 kHz.
The abnormality diagnosis device according to any one of claims 1 to 4. The abnormality determination means has a means for calculating the degree of abnormality of the frequency band by calculating the degree of matching with the predetermined reference data of the corresponding frequency band for each of the data classified by frequency band,
The system further comprises a second history display control unit that performs control to display the degree of abnormality historically.
The abnormality diagnosis device according to any one of claims 1 to 5. The second history display control means identifies and displays the degree of abnormality in the display of the history.
The abnormality diagnosis device according to claim 6. Computer,
Analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle,
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
An abnormality determination unit that determines presence or absence of an abnormality of the frequency band by determining whether each of the frequency band classified data conforms to predetermined reference data of the corresponding frequency band,
Abnormality cause determining means for determining an abnormality cause determined based on the presence or absence of an abnormality in each of the frequency bands determined by the abnormality determining means;
Program to function as. Computer,
Analysis means for performing octave band analysis of vibration data detected by a vibration sensor installed on a railway vehicle,
Classification means for classifying the analysis result of the analysis means into frequency band data of at least three frequency bands of low frequency band, middle frequency band and high frequency band;
An abnormality determination unit that determines presence or absence of an abnormality of the frequency band by determining whether each of the data according to frequency band conforms to predetermined reference data of the corresponding frequency band at a predetermined cycle,
A first control is performed to historically display a ratio of the number of times that the abnormality determination unit determines that there is an abnormality during a predetermined unit time or a predetermined travel unit longer than the predetermined period for each of the frequency bands. History display control means,
Program to function as.

Images