JPWO2018138862A1 - State judgment system - Google Patents

Abstract

The state determination system according to the present invention includes an imaging unit that images a waveform generated in response to an operation of a state determination target, and a reference image storage that stores a plurality of images of the waveform in the reference state of the state determination target as a reference image. Means for comparing the inspection image data with a plurality of data of the reference image using the waveform image obtained when the state determination is performed as an inspection image, and between the inspection image data and the reference image data. Collating means for classifying the differences, and determining means for determining that the difference is not significant based on the grade calculated by the collating means and that the difference is not significant and that the difference is not significant. Since the data of the reference image obtained by imaging the waveform and the data of the inspection image are compared, the determination by the arithmetic processing is possible without depending on the person who has the skill to make the determination.

Description

  The present invention relates to a state determination system that determines whether a state determination target is in a planned state (normal state) or deviated from the planned state (abnormal state).

  When a large number of devices are installed and used on a large site such as a factory, these devices are usually centrally managed by a control center such as a management center where various information indicating the operating state is collected. Yes.

  In this case, as data collected on the control side, for example, information on voltage and current is collected for electric devices, and information on flow rate and temperature is collected for devices that handle fluids (pumps, compressors, etc.) Centralized to the control side. Then, it is determined by the control processing means on the control side whether a predetermined state is maintained through comparison processing such as a threshold value.

  However, in the information indicating the operation state of the apparatus / equipment, even if the information is processed by a known calculation method, whether or not the planned state is maintained based on the result obtained from the information is determined. Some things were difficult. For example, a waveform generated by an operation of a state determination target, such as a signal waveform obtained by an analog oscilloscope or a digital oscilloscope, is useful information for determining the state of a determination target, but includes many determination elements and is simple. It is difficult to make an accurate determination by calculation processing based on comparison with a threshold value. For this reason, the state determination using waveforms relies on the skill of a skilled determiner.

  On the other hand, the determination based on the skill of the determiner cannot be performed unless there is a person who has the skill to make the determination at a place where the waveform can be confirmed. Therefore, a method has been proposed that enables state determination even when there is no person who has the skill to make a determination at a place where the waveform can be confirmed. For example, in Japanese Patent Laid-Open No. 2003-256028, each signal waveform of a surface inspection machine can be easily observed from a remote place with a terminal personal computer, and failure analysis by a specialist (person who has a skill capable of making a determination) is performed. There has been proposed a remote monitoring system that can be performed quickly.

Japanese Patent Laid-Open No. 2003-256028

  However, even if the waveform can be confirmed at a remote place, it is difficult for a person having the skill to make a determination to continue monitoring the waveform. Therefore, in the conventional state determination method that performs state determination using the waveform by the skill of the determiner, there is a possibility that the normal state suddenly becomes an abnormal state, or the normal state (scheduled state) It is difficult to apply when an abnormal state (a state deviating from a planned state) is repeated.

  Therefore, an object of the present invention is to provide a state determination system that enables continuous monitoring of a state determination target using a waveform generated with the operation of the state determination target.

  The state determination system according to the present invention includes an imaging unit that images a waveform generated in response to an operation of a state determination target, and a reference image storage that stores a plurality of images of the waveform in the reference state of the state determination target as a reference image. Means for comparing the data of the inspection image and a plurality of data of the reference image using the image of the waveform obtained when the state determination is performed as an inspection image, and between the data of the inspection image and the data of the reference image Collating means for grading differences, and determining means for determining that the difference is not significant based on the grade calculated by the collating means and that the difference is not significant and that the difference is not significant.

  In the present invention, the fact that the difference is significant means that the difference exceeds an allowable range. For example, if there is a range in the state quantity obtained in the normal state of the device to be monitored, it can be accepted if the difference in the data is within the range of the range, so both data match, not a significant difference It is judged. On the other hand, when the difference in data is out of the range of the state quantity obtained in the normal state, it is determined that the difference is significant and the two data do not match.

  The waveform may include a change over time of a physical quantity that changes according to a movement of a driving part of the state determination target. In addition, the physical quantity that changes depending on the movement of the drive part is, for example, vibration, pressure, sound, acceleration, tension, heat quantity, flow rate, light quantity, deformation quantity, and odor quantity.

  The waveform may include a change in current or voltage with time in power supply to the state determination target.

  The waveform may include a waveform of a signal received or transmitted by the state determination target.

  The waveform imaging may be performed on a plurality of types of waveforms having different properties.

  The reference image and the inspection image may be constituted by data obtained by extracting the waveform data for a predetermined period.

  The grade may be written on the inspection image and delivered to the determination unit.

  The determination means determines the presence or absence of an abnormal state based on the determination as to whether or not the reference state, and when the cumulative number of occurrences of the abnormal state data reaches a predetermined number, or the frequency of occurrences is a predetermined number When reaching the above, it may be determined that the lifetime of the state determination target.

  In the present invention, in the reference state of the state determination target, the image of the reference image obtained by imaging the waveform generated by the operation of the state determination target and the waveform generated by the operation of the state determination target when performing the state determination Since the inspection image data is compared, the determination by the arithmetic processing is possible without depending on the person who has the skill to make the determination.

  Further, the difference between the inspection image data and the reference image data is classified, and based on this grade, it is determined that the reference state is not significant and the reference state is not significant when the difference is not significant. Therefore, more accurate determination can be performed as compared with the conventional arithmetic processing by comparison with a simple threshold value.

  Furthermore, since a plurality of waveform images in the reference state of the state determination target are used as the reference image, the determination accuracy can be improved.

  Therefore, according to the present invention, the state determination can be performed by the arithmetic processing using the waveform generated in accordance with the operation of the state determination target, and the state determination target can be continuously monitored.

It is a block diagram of the state determination system which concerns on this invention. It is a functional block diagram of a sensor controller. It is a functional block diagram of a storage server and a search engine server.

An embodiment of an apparatus / equipment abnormality detection system according to the present invention will be described.
This apparatus / equipment abnormality detection system is configured to detect these abnormalities remotely using a rotating apparatus as a state determination target 10 in a factory where a plurality of rotating apparatuses such as motors are installed. In this embodiment, the plurality of state determination targets 10 are collectively controlled by the PLC 5 connected to the communication line 9, but may be controlled by a dedicated controller.

  The state determination target 10 is provided with a sensor 1 that detects vibration as a state quantity that changes depending on the operation. In this embodiment, a pair of axial vibration sensors is used as the sensor 1. One of the pair of sensors 1a and 1b (first sensor 1a) is installed in a direction for detecting the axial vibration vertical component of the state determination target 10, and the other (second sensor 1b) is installed in the direction for detecting the axial vibration horizontal component. Has been.

  The sensor 1 is not limited as long as it can detect a state quantity that changes depending on the operation state of the state determination target 10 and is necessary for generating an image of a waveform generated in accordance with the operation of the state determination target 10. A pressure sensor, an acceleration sensor, a rotation speed sensor, a sound sensor, a torque sensor, a load current sensor, or the like may be used. Further, they may be used alone, two may be used as a pair, or a plurality may be used as a set. When a pair or a plurality are used as a set, different types of combinations may be used. For example, as a combination of the first sensor 1a and the second sensor 1b, a combination of a pressure sensor and a vibration sensor, or a combination of an acceleration sensor and a voice sensor may be employed.

  The sensor 1 is connected to the sensor controller 2. The sensor controller 2 is connected to the search engine server 3 and the storage server 4 via the communication line 9. The search engine server 3 corresponds to determination means of the present invention, and vibration data obtained by the sensor 1 is converted into sensor signal waveform image data by the sensor controller 2 and transmitted to the search engine server 3 via the communication line 9. And the determination process described later is performed.

<Sensor controller>
As shown in FIG. 2, the sensor controller 2 includes a first data conversion unit 21a, a second data conversion unit 21b, an image data transmission unit 22, a communication unit 23, a storage unit 24, a mode determination unit 25, and a reference image generation switch 26. Is provided.

  The first data conversion means 21a constitutes the imaging means of the present invention together with the first sensor 1a, and the vibration data obtained via the first sensor 1a is converted into waveform image data (sensor signal waveform indicating the change over time in amplitude). Image data), data for one frame period (hereinafter referred to as “frame image data”) is extracted and delivered to the image data transmitting means 22. Similarly, the second data conversion means 21b constitutes the imaging means of the present invention together with the second sensor 1b, converts vibration data obtained via the second sensor 1b into sensor signal waveform image data, and generates frame image data. Is delivered to the image data transmission means 22. One frame period corresponds to the predetermined period of the present invention, but may be set as appropriate according to the detection target and detection conditions.

  The image data transmission unit 22 performs the following processing according to the mode data output from the mode determination unit 25. First, when the mode data indicates the online mode, the frame delivered from the first data conversion means 21a to the sensor No. including the device type assigned to the sensor controller 2 and the sensor type data assigned to the first sensor 1a. Image data is assigned and delivered to the communication means 23. Further, the frame image data handed over from the second data conversion means 21b is assigned to the sensor No. including the device No. given to the sensor controller 2 and the sensor type data given to the second sensor 1b, and the communication means 23 is given. hand over.

  In the online mode, the waveform image composed of the frame image data delivered to the communication unit 23 corresponds to the inspection image of the present invention.

  On the other hand, when the mode data indicates the reference image generation mode, one or both of the frame image data delivered from the first data conversion unit 21a or the second data conversion unit 21b is used as an identifier indicating the reference image. And give it to the communication means 23.

  The reference image only needs to be a reference for determining whether or not the state determination target 10 is abnormal. For example, data obtained during the operation of a malfunctioning apparatus of the same model as the state determination target 10 is employed. Alternatively, data obtained when the state determination target 10 is operating normally may be used.

  The communication unit 23 transmits the data delivered from the image data transmission unit 22 to the search engine server 3 and the storage server 4 via the communication line 9. In addition, when it is determined that the state determination target 10 is in an abnormal state by the state determination described later and the abnormal state data transmitted from the determination device 6 is received, the abnormal state data is delivered to the storage unit 24.

  The storage unit 24 includes a device No. assigned to the sensor controller 2, a sensor No. including sensor type data given to the first sensor 1a and the second sensor 1b, and a reference image No. suitable for the state determination target 10. Is stored and transferred to the image data transmission means 22. Further, when the abnormal state data is delivered from the communication means 23, it is stored.

  As a method for inputting the device No., the sensor No. including the sensor type data, and the reference image No. stored in the storage unit 24, a known method suitable for the shape and installation state of the sensor controller 2 may be appropriately employed. For example, it may be set using a separate terminal device that exchanges data without contact with the sensor controller 2.

  The mode determination unit 25 outputs mode data corresponding to the mode selected by operating the reference image generation switch 26 to the image data transmission unit 22. In this embodiment, the reference image generation switch 26 is always OFF, and the reference image generation mode is selected by turning it ON. When the reference image generation switch 26 is turned on, mode data indicating the reference image generation mode is output to the image data transmission unit 22. On the other hand, at all times (when the reference image generation switch 26 is OFF), mode data indicating the online mode is output to the image transmission means 22.

<Search engine server>
The search engine server 3 includes a communication unit 31 and a collation unit 32 as shown in FIG.

  The communication unit 31 passes the data transmitted from the sensor controller 2 to the verification unit 32. In addition, data that matches the request of the verification unit 32 is acquired from the storage server 4 via the communication line 9, and is transferred to the verification unit 32. Further, the frame image data handed over from the collating means 32 and written with a numerical value of similarity described later is handed over to the determination device 6 via the communication line 9.

  The collation unit 32 collates the inspection image data with the reference image data. First, one image No. is determined from a plurality of reference images based on device No data, sensor No data including sensor type data, and reference image No included in data delivered from the communication unit 31. Then, a reference image data acquisition request is output to the communication unit 31. Further, the frame image data included in the data delivered from the communication unit 31 is temporarily stored as inspection image data until the collation process is completed.

  The communication unit 31 that has received the data acquisition request acquires data matching the request, that is, the reference image data from the storage server 4 via the communication line 9 and passes it to the verification unit 32 as described above. .

  The collation unit 32 that has received the reference image data from the communication unit 31 uses the frame image data included in the data delivered from the communication unit 31 as inspection image data, and the inspection image data and the reference image data based on each method. Perform the verification.

  Image matching can be performed by matching pattern elements and features extracted from image data between two parties based on each method. Further, the image data has parameters such as position and direction, and in image collation, the arrangement in the parameter space is collated as a pattern.

  In this embodiment, template matching is employed as an image matching method. The reference image data is referred to as a state determination target 10 template.

  In the template matching method, a portion similar to the reference image data is detected in the inspection image data. Specifically, the similarity (or dissimilarity) of the overlapping regions is calculated while sequentially shifting the reference image data on the inspection image data. It shows that the image of the part which overlaps with the reference image data which is a template is similar, so that the calculated value of similarity is small. Further, when they completely match, the calculation result value is 0, and the smaller the value close to 0, the more similar.

  In the case of similarity at an arbitrary position of the inspection image data, if the value of the above calculation becomes smaller than a certain level, this position is specified as a detection position.

  In this embodiment, the cross-correlation method is employed as the similarity calculation method, but since it is a known method, a detailed description of the calculation formula and the like is omitted.

  The numerical value of the predetermined similarity obtained as the calculation result value is written and processed at a position specified in the frame image data stored as the inspection image. Then, a data acquisition request for the next reference image is output to the communication unit 31.

  In this embodiment, the numerical value of the similarity degree itself is treated as a grade, but the grade may be defined within a predetermined numerical range.

  The collation unit 32 that has received the data of the next reference image from the communication unit 31 repeats the above collation until the collation process for all of the plurality of reference images is completed. When the collation processing for all the reference images is completed, the frame image data in which the numerical values of the similarity of each of the plurality of reference images are written is delivered to the determination device 6 via the communication unit 31.

  In addition to the pattern matching described above, the image matching method includes DP matching, which is a one-dimensional elastic matching technique, and graphs that use features as vertices and the relationships between features as edges, and perform correlation between the graphs. Thus, there is a graph matching for matching feature sets. You may employ | adopt these methods according to a use condition.

  In the determination device 6 that has received the frame image data in which the similarity values of each of the plurality of reference images are written, the state determination of the state determination target 10 is performed by the internal determination unit 61 based on the similarity values. I do.

  First, the determination unit 61 determines that the difference between the inspection image and the reference image is significant when the degree of similarity written in the frame image data includes a predetermined threshold value or more. If all the similarities are equal to or less than the threshold, it is determined that the difference between the inspection image and the reference image is not significant.

  The determination of whether or not an abnormal state is present depends on the nature of the reference image. For example, when the data on the failure state of the device that is the state determination target 10 can be acquired, the waveform image obtained in the failure state can be used as the reference image. In this case, the difference between the inspection image and the reference image is not significant. The determination that the reference state is not present means an abnormal state of the state determination target 10.

  On the other hand, when data on the failure state of the device that is the state determination target 10 cannot be acquired, the waveform image obtained in the normal state can be used as the reference image. In this case, the difference between the inspection image and the reference image is significant. Determination that the state is not the reference state means an abnormal state of the state determination target 10.

  In addition, when failure state data can be acquired, determination using both a failure state waveform image and a normal state waveform image may be performed. For example, the normal state is the first reference state, the normal state waveform image is the first reference image, the failure state is the second reference state, the failure state waveform image is the second reference image, and the first reference image is first When it is determined that the current state is not the first reference state, the second reference image is compared. And when it determines with it being a 2nd reference | standard state, you may determine with it being in an abnormal state.

  The determination means 61 determines the presence / absence of an abnormal state of the state determination target 10 based on determination of whether or not the state determination target 10 is in the reference state based on these criteria. If the determination unit 61 determines that the state is abnormal, the determination device 6 uses the frame image data in which the numerical value of the similarity is written as abnormal state data, and the sensor controller 2 of the state determination target 10 and the storage server. 4 through the communication line 9.

  As shown in FIG. 3, the storage server 4 includes a communication unit 41 and a storage unit 42.

  The communication unit 41 receives the reference image data transmitted from the sensor controller 2 and delivers it to the storage unit 42. In response to a request from the search server 3, basic image data is extracted from the data stored in the storage unit 42 and transmitted to the search engine server 3 via the communication line 9. Further, the abnormal state data transmitted from the determination device 6 is received and delivered to the storage means 42.

  The storage means 42 accumulates the reference image data delivered from the communication means 41 and the abnormal state data.

  The data stored in the storage means 42 can be browsed by the determination device 6. In this case, the communication unit 41 that has received the browsing request transmitted from the determination device 6 extracts data that meets the request from the data stored in the storage unit 42, and delivers it to the determination device 6 via the communication line 9.

  In this embodiment, the search engine server 3 and the storage server 4 are separate, but may be an integrated server 30 as indicated by phantom lines in FIGS.

  The determination device 6 is not limited as long as it includes an arithmetic processing unit capable of determining the presence / absence of the abnormal state described above, and a known arithmetic processing device can be employed. Further, more detailed determination regarding the abnormal state may be performed according to the performance of the arithmetic processing.

  For example, the life determination of the state determination target 10 may be performed using the above-described determination of the presence or absence of an abnormal state. When the cumulative number of occurrences of abnormal state data with respect to the state determination target 10 reaches a predetermined number, or when the frequency of the number of occurrences reaches a predetermined number, it is determined that the state determination target 10 is at the end of its life and prompts an alarm. It may be a thing.

In the embodiment described above, shaft vibration is adopted as a condition for determining the presence / absence of an abnormal state. However, a plurality of viewpoints may be used for determining the presence / absence of an abnormal state. For example, air pressure and voltage, sound and temperature may be used as the determination conditions.
Multiple parameter combinations can be determined in advance

  Furthermore, a plurality of state determination targets 10 that are related may be selected in advance, and the presence / absence of an abnormal state may be determined based on information regarding the plurality of state determination targets 10.

  In the above embodiment, the time-dependent change of the physical quantity (time-dependent change of the axial vibration) that changes due to the movement of the drive part of the state determination target 10 is adopted as the waveform generated by the operation of the state determination target 10. In the case of an apparatus having a communication function, a waveform of a signal that is received or transmitted by the state determination target may be employed.

  When a transmission signal transmitted / received by the state determination target is affected by noise or attenuated, data transmission / reception as intended is not performed. Therefore, when normal data transmission / reception is performed, the signal waveform when the signal is affected or attenuated differs from the normal signal waveform. . Therefore, it is possible to determine whether the signal is in a normal state based on the difference in the signal waveform.

DESCRIPTION OF SYMBOLS 1 Sensor 1a 1st sensor 1b 2nd sensor 2 Sensor controller 3 Search engine server 4 Storage server 5 PLC
6 determination device 9 communication line 10 state determination target 21a first data conversion means 21b second data conversion means 22 image data transmission means 23, 31, 41 communication means 24, 42 storage means 25 mode determination means 26 reference image generation switch 30 Body server 32 Verification means 61 Determination means

Claims (8)

  When performing state determination, imaging means for imaging a waveform generated by the operation of the state determination target, reference image storage means for storing a plurality of images of the waveform in the reference state of the state determination target as a reference image, and A collating means for comparing the data of the inspection image with a plurality of data of the reference image using the obtained waveform image as an inspection image, and grading the difference between the data of the inspection image and the data of the reference image; A state determination system comprising: determination means for determining that a reference state is not significant when the difference is not significant based on the grade calculated by the collating means, and that the difference is not a reference state when significant.   The state determination system according to claim 1, wherein the waveform includes a change over time of a physical quantity that changes according to a movement of a driving part of the state determination target.   The state determination system according to claim 1, wherein the waveform includes a change over time in current or voltage in power supply to the state determination target.   The state determination system according to claim 1, wherein the waveform includes a waveform of a signal received or transmitted by the state determination target.   The state determination system according to claim 1, wherein the waveform imaging is performed for a plurality of types of waveforms having different properties.   The state determination system according to claim 1, wherein the reference image and the inspection image include data obtained by extracting the waveform data for a predetermined period.   The state determination system according to claim 1, wherein the grade is written on the inspection image and delivered to the determination unit.   The determination means determines the presence or absence of an abnormal state based on the determination as to whether or not the reference state, and when the cumulative number of occurrences of the abnormal state data reaches a predetermined number, or the frequency of occurrences is a predetermined number The state determination system according to any one of claims 1 to 7, wherein the state determination target is determined to be the life of the state determination target.

Images