JP2016090461A - Construction method for abnormality diagnostic functional model for sound of monitored object, and construction system therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a construction method for an abnormality diagnostic functional model for a sound of monitored object that finds regularity of a sound (vibration) when the monitored object is abnormal through basic experiment, and establishes a basic part for establishing a high-precision inspection method, and a construction system therefor.SOLUTION: A construction method for an abnormality diagnostic functional mode for a sound of a monitored object comprises: (a) clarifying a sound sense factor group as elements capable of illustrating differences between sounds (vibration) of the monitored object generated when the monitored object is normal and abnormal; (b) combining physical characteristics and a model for sound (vibration) of a person together that a machine can observe, the sound sense factor group of the sound (vibration) of the monitored object being a model for a sound (vibration) that the person senses; (c) observing and illustrating the differences between the sounds (vibration) when the monitored object is normal and abnormal with the sound sense factor group; and (d) clarifying whether the differences between the sounds (vibration) when the monitored object is normal and abnormal as mentioned above can be illustrated with a physical characteristic group, and illustrating the same with the sound sense factor group and physical characteristic group so as to make correspondence between both clear.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an abnormality detection method and system imitating sensory inspection of a monitoring target, and more particularly to a method and system for constructing an abnormal diagnosis sensory model of a sound to be monitored.

  Conventionally, for example, a railway vehicle state monitoring system has been proposed as shown in Patent Document 1 and Non-Patent Documents 1 to 3 below.

JP 2011-51518 A

Nakae S.M. et al. , Dassen kenchi souchi no shiken (in Japan), Railway Technical Research Pre-Report, (1982) Hayashi Y. et al. et al. , Condition Monitoring and Fault Detection of Railway Vehicle (in Japan), Translog 2007 symposium, pp. 331-334 (2007) Morikawa M.M. et al. , Study on Detection of Signs of Weelcrimb Delament (in Japan), J-Rail 2008 Symposium (2008).

  A malfunction of a traveling device such as a railway vehicle to be monitored may lead to a major accident. In the inspection of railway cars such as Shinkansen, in addition to tightening torque management with a dedicated tool, the head of the bolt called a hammering test is struck with a dedicated hammer, and the inspector checks whether the bolt is loose by the sound at that time. Is also used. In this method, an experienced inspector uses various information such as sound attenuation as well as sound level to judge a defect. Compared to this, the abnormality detection method using a machine relies only on the amplitude for each frequency. Here, even in machine inspections, if it is possible to make multiple judgments using various information such as sound attenuation like human sensory inspections, machine detection can be expected more than ever, further improving safety. Improvement can be expected. Therefore, in the present invention, the basic part for establishing a highly accurate inspection method is found by finding the regularity of the sound (vibration) at the time of abnormality by a basic experiment or the like.

  The present invention provides a method for constructing a sound abnormality diagnostic sensory model in which the abnormality diagnosis sensory model of the sound to be monitored is taken into account in the above-described state monitoring of the railway vehicle, and is integratedly determined to improve the accuracy of the abnormality diagnosis, and The purpose is to provide such a system.

In order to achieve the above object, the present invention provides
[1] In the construction method of the abnormal diagnosis sensory model of the sound to be monitored,
(A) Clarify a group of sensation factors that can explain the difference in the sound (vibration) of the monitoring target when the monitoring target is normal and abnormal,
(B) The sound factor group of the sound (vibration) to be monitored is a model of sound (vibration) that is perceived by humans, and the physical characteristics that can be observed by the machine and the human sound ( Vibration) model,
(C) Observe the difference in sound (vibration) between the normal state and the abnormal state of the monitoring target, and explain the sound factor group of sound (vibration);
(D) It will be clarified how the difference in sound (vibration) between the normal state and the abnormal state of the monitoring target can be explained by the physical characteristic group, and the same thing will be explained by the sound factor group and the physical characteristic group. The feature is to clarify the correspondence between the two.

  [2] In the method of constructing the abnormal diagnosis sensory model for the sound to be monitored as described in [1] above, the difference in sound (vibration) between the normal state and the abnormal state of the monitored object is explained in terms of physical characteristics group, and clearly the same An object is explained by a sound factor group and a physical property group, and the correspondence between the two is performed.

  [3] In the construction method of the abnormality diagnostic sensory model for the monitoring target sound described in [1], the monitoring target is a loose bolt.

  [4] In the method for constructing the abnormality diagnosis sensory model of the sound to be monitored as described in [2] above, “high sound” is generated when the bolt is tightened, and “dull sound” is generated when the bolt is loose. Therefore, since a change in tightening torque appears in a change in sound, the torque is determined based on the degree of change in the sound, and the relationship between the change in sound and the tightening torque is investigated.

  [5] In the construction method of the abnormal diagnosis sensory model of the sound to be monitored as described in [1] above, the physical characteristics of the sound (vibration) are frequency, frequency fluctuation, amplitude fluctuation, attenuation, sound amplitude ratio, harmonic overtone. It is a sound other than.

  [6] In the system for constructing the abnormal diagnosis sensory model for the sound to be monitored, the abnormality diagnosis sensory model for the sound to be monitored is taken into consideration and integrated judgment is performed to improve the accuracy of the abnormality diagnosis.

  [7] In the system for constructing an abnormal diagnosis sensory model for a sound to be monitored as described in [6] above, the monitoring object is loosening of a bolt.

  [8] In the system for constructing the abnormality diagnosis sensory model for the sound to be monitored according to [6] above, the monitoring object is wheel wear.

  [9] In the system for constructing an abnormal diagnosis sensory model for a sound to be monitored as described in [6] above, the monitoring object is a crack in a frame.

  ADVANTAGE OF THE INVENTION According to this invention, the abnormality diagnosis sensory model of the sound to be monitored can be considered and integrated judgment can be performed, and the accuracy of abnormality diagnosis can be improved.

It is a block diagram which shows the abnormality diagnosis sensory model of the bolt fastening sound in a railway vehicle state monitoring system.

The method for constructing the abnormal diagnosis sensory model of the sound to be monitored of the present invention is as follows.
(A) Clarify a group of sensation factors that can explain the difference in the sound (vibration) of the monitoring target when the monitoring target is normal and abnormal,
(B) The sound factor group of sound (vibration) to be monitored is a model of sound (vibration) that is sensed by humans. Linking physical properties that can be used with human sound (vibration) models,
(C) Observe the difference in sound (vibration) between normal and abnormal, and explain with sound (vibration) factor groups,
(D) It will be clarified how the difference in sound (vibration) between the normal time and the abnormal time can be explained by the physical characteristic group, and the same thing will be explained by the sound factor group and the physical characteristic group.

  Hereinafter, embodiments of the present invention will be described in detail.

The present invention is intended for bolt loosening, wheel wear, cracks in the frame, and the like. Here, bolt loosening will be described as an example.
1. Construction of sensory model for abnormality diagnosis of sound (vibration) to be monitored 1) List of factor groups of sound (vibration) to be monitored Sound sense that can explain the difference between sound (vibration) when monitoring target is normal and abnormal Clarify factor groups. For example, words that express the difference between normal sound and abnormal sound expressed in words such as loud sounds and dull sounds are arranged.
2) Enumeration of physical characteristics groups Sound (vibration) factor groups are models of sounds (vibrations) that are perceived by humans. In order to apply them to machine abnormality diagnosis, machines must be observed. It is necessary to link physical characteristics that can be used with human sound (vibration) models. Therefore, the physical characteristics of sound (vibration) are enumerated, and we will investigate which of these will be used while conducting research.

  Examples of physical characteristics of sound (vibration) to be monitored are given below.

  a. Frequency b. Frequency fluctuation c. Amplitude fluctuation d. Attenuation e. Overtone amplitude ratio f. Sound other than harmonics.

(3) Connection of sound (vibration) factor group and physical property group The difference between normal and abnormal sound (vibration) will be observed and explained with the sound (vibration) factor group. Next, it will be clarified how the difference in sound (vibration) between the normal time and the abnormal time can be explained by the physical property group. Since the same thing was explained with the sound factor group and the physical property group from these, the correspondence between the two can be clarified.

(4) Case study in normal and abnormal cases In the above example, when the bolt is tightened, a “high sound” is heard, and when the bolt is loose, a “blunt sound” is heard. Appears in the change of sound, and this degree can be used to guess the tightening torque. For this purpose, the relationship between sound change and tightening torque is investigated.

  In this way, an index using a plurality of physical property groups is completed using the sensual pitch factor group.

The following is a specific explanation.
FIG. 1 is a block diagram illustrating a sensor model for abnormality diagnosis of bolt tightening noise in a railway vehicle state monitoring system.

  In this figure, the upper column (a) is a column indicating the state of the monitoring target. For example, when the bolt is tightened, the middle column (b) is a column indicating the factor of sound (vibration). When the bolt is tightened, the sound (vibration) factor is a high sound. The lower column (c) is a column indicating a physical characteristic group of sound (vibration). For example, in a state where the bolt is tightened, the degree of relation with the frequency of the sound (vibration) is +1.0, and the attenuation is -1.0.

  Here, the degree of relevance is obtained by striking a hammer with a hammer when the bolt is tightened or loosened, and comparing the sound of the two with respect to each physical characteristic. For example, at a high frequency, the tightened state is high, and when it is loose, it is low. Therefore, the frequency of the tightened state is +1.0, and the frequency of the loose state is −1. . Score 0.

  On the other hand, when the bolt is loose, the middle column (b) has a dull sound, and the lower column (c) has a frequency height of -1.0 and an attenuation of +1.0. Note that this is a column indicating a physical property group of sound (vibration). For example, in a state where the bolt is tightened, the frequency of the sound (vibration) is +1.0, and the attenuation is -1.0.

  In addition to the above, in the column indicating the physical property group, frequency fluctuations, overtone amplitude ratios, sounds other than overtones, and the like can be listed.

  Thus, even in the state monitoring of whether the bolt is tightened or loosened, it is possible to monitor the state with high accuracy by constructing a sound abnormality diagnosis sensory model that comprehensively determines physical characteristics. .

  In addition, this invention is not limited to the said Example, Based on the meaning of this invention, a various deformation | transformation is possible and these are not excluded from the scope of the present invention.

  The method and system for constructing an abnormal diagnosis sensory model of a sound to be monitored according to the present invention, including the abnormal diagnosis sensory model of the sound to be monitored, integrated judgment, and improved accuracy of abnormality diagnosis It can be used as a construction method and a construction system of a sensory abnormality diagnosis sensory model.

Claims (9)

(A) Clarify a group of sensation factors that can explain the difference in the sound (vibration) of the monitoring target when the monitoring target is normal and abnormal,
(B) The sound factor group of the sound (vibration) to be monitored is a model of sound (vibration) that is perceived by humans, and the physical characteristics that can be observed by the machine and the human sound ( Vibration) model,
(C) Observe the difference in sound (vibration) between the normal state and the abnormal state of the monitoring target, and explain the sound factor group of sound (vibration);
(D) It will be clarified how the difference in sound (vibration) between the normal state and the abnormal state of the monitoring target can be explained by the physical characteristic group, and the same thing will be explained by the sound factor group and the physical characteristic group. A method for constructing a sensory model for abnormal diagnosis of sound to be monitored, characterized by clarifying the correspondence between the two.   2. The method of constructing a sensory abnormality model for a sound to be monitored according to claim 1, wherein the difference in sound (vibration) between the normal state and the abnormal state of the monitoring target is described in terms of physical characteristics, and clearly the same thing A method for constructing an abnormal diagnosis sensory model of a sound to be monitored, characterized in that the group and the physical characteristic group are explained and the two are dealt with.   2. The method for constructing an abnormal diagnosis sensory model for a monitoring target sound according to claim 1, wherein the monitoring target is a bolt loosening.   3. The method for constructing a sensory abnormality sensory model for a sound to be monitored according to claim 2, wherein a "high sound" is produced when the bolt is tightened and a "blunt sound" is produced when the bolt is loose. Because the change in the tightening torque appears in the change in the sound, the torque is judged based on the degree of the change in the sound, and the relationship between the change in the sound and the tightening torque is investigated. How to build a sensory model.   2. The method for constructing an abnormal diagnosis sensory model of a sound to be monitored according to claim 1, wherein the physical characteristics of the sound (vibration) are frequencies other than frequency, frequency fluctuation, amplitude fluctuation, attenuation, sound amplitude ratio, and harmonics. A method for constructing a sensory model for abnormal diagnosis of sound to be monitored, characterized by being.   In the system for building sensory abnormality sensory model for monitoring target sound, the abnormality of monitoring target sound is characterized by improving the accuracy of abnormality diagnosis by taking into account the sensory abnormality model of monitoring target sound and making an integrated judgment A diagnostic sensory model construction system.   7. The system for constructing an abnormal diagnosis sensory model for a monitoring target sound according to claim 6, wherein the monitoring target is a bolt loosening.   The system for constructing an abnormal diagnosis sensory model for a monitoring target sound according to claim 6, wherein the monitoring target is wheel wear.   7. The system for constructing an abnormal diagnosis sensory model for a sound to be monitored according to claim 6, wherein the monitor object is a crack in a frame.

Images