JP2018092437A - Production facility abnormality diagnosis system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a production facility abnormality diagnosis system capable of accurately determining the presence or absence of an abnormality in a production facility.SOLUTION: A production facility abnormality diagnosis system comprises item group acquisition means 10, unit space creation means 110, Mahalanobis distance calculation means 111, and determination means 112. The item group acquisition means 10 acquires an item group indicating a state of a production facility 2 at every predetermined time t2. The unit space creation means 110 creates a unit space on the basis of the acquired item group. The Mahalanobis distance calculation means 111 calculates a Mahalanobis distance MD on the basis of the acquired item group and the unit space. The determination means 112 determines that a state of the production facility 2 is abnormal when the calculated Mahalanobis distance MD exceeds a threshold value MDa. The unit space creation means 110 creates a unit space on the basis of an item group excluding the item group determined to be abnormal by the determination means 112, among the item groups acquired by the item group acquisition means 10.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a production facility abnormality diagnosis system.

  Patent Document 1 discloses a plant state monitoring device that monitors a plant state using an MT (Mahalanobis Taguchi) method. This plant condition monitoring unit is based on the state quantity of the plant in the period from the time point retroactive to the past for the first period to the time point retroactive to the second period in the past. Create a space (reference space). Then, the Mahalanobis distance is calculated based on the created unit space and the plant state quantity acquired when the plant state is evaluated, and when the Mahalanobis distance exceeds the threshold, the plant state is determined to be abnormal. In this MT method, even when there are a large number of state quantities to be monitored in the plant, it is possible to easily diagnose a plant abnormality.

Japanese Patent No. 5031088

  By the way, when the plant state monitoring data includes data at the time of abnormality in the plant state quantity data used for creating the unit space, the Mahalanobis distance is based on the state quantity of the plant including the data at the time of abnormality. Since it is calculated, it cannot be correctly determined whether there is an abnormality in the plant state.

  This invention is made | formed in view of the said situation, Comprising: It aims at providing the production facility abnormality diagnosis system which can determine correctly the presence or absence of abnormality of a production facility.

  In order to achieve the above object, a production facility abnormality diagnosis system according to one aspect of the present invention is a production facility abnormality diagnosis system that diagnoses an abnormality in a production facility using the MT method, and has the following features. The production facility abnormality diagnosis system includes item group acquisition means, unit space creation means, Mahalanobis distance calculation means, and determination means. The item group acquisition unit acquires an item group including a plurality of items indicating the state of the production facility at predetermined time intervals. The unit space creating means creates a unit space based on the item group acquired by the item group acquiring means. The Mahalanobis distance calculating means calculates the Mahalanobis distance based on the item group acquired by the item group acquiring means and the unit space created by the unit space creating means. The determination means determines that the state of the production facility is abnormal when the Mahalanobis distance calculated by the Mahalanobis distance calculation means exceeds a preset threshold. The unit space creating unit creates the unit space based on the item group excluding the item group determined to be abnormal by the determining unit among the item group acquired by the item group acquiring unit.

  The production facility abnormality diagnosis system according to one aspect of the present invention can accurately determine the presence or absence of abnormality in the production facility.

FIG. 1 is an explanatory diagram showing a production facility diagnosed by a production facility abnormality diagnosis system according to an embodiment of the present invention. FIG. 2 is a block diagram of the above-described production facility abnormality diagnosis system. FIG. 3 is a flowchart of the above-described production facility abnormality diagnosis system. FIG. 4 is an explanatory diagram showing a method for creating a unit space by the production facility abnormality diagnosis system described above.

  The embodiment described below relates to a production facility abnormality diagnosis system, and more particularly to a production facility abnormality diagnosis system that diagnoses the presence or absence of abnormality in a production facility using the MT method.

(Production equipment)
The production facility abnormality diagnosis system of the present embodiment diagnoses whether there is an abnormality in the production facility 2 shown in FIG. The production facility 2 continuously produces a large number of exterior materials. The exterior material produced by the production facility 2 is a flat roof material called flat slate tile, and is a cured product of a hydraulic material mainly composed of cement.

  As shown in FIG. 1, the production facility 2 includes a conveyor 20, a water feeder 21, a hydraulic material feeder 22, a watering machine 23, a cutting machine 24, a cement material spreader 25, a chromite spreader 26, and a press machine 27. I have.

  The conveyor 20 of this embodiment is a belt conveyor. The water supply device 21, the hydraulic material supply device 22, the watering device 23, the cutting device 24, the cement material spreading device 25, and the coloring stone spreading device 26 are arranged in this order from the upstream side in the conveying direction of the conveyor 20.

  The press machine 27 includes a material spreading thickness adjusting roll 270 and a plurality of pressure rolls 271 to 274 provided above the conveyor 20. Specifically, in the conveying direction of the conveyor 20, the press machine 27 of the present embodiment is between the hydraulic material supply machine 22 and the watering machine 23, between the watering machine 23 and the cutting machine 24, and between the cement material spreading machine 25 and The pressure rolls 271 to 274 are provided between the Aishi stone spreader 26 and downstream of the Aya stone spreader 26.

  The water feeder 21 supplies water onto the conveyor 20.

  The hydraulic material supplier 22 supplies the hydraulic material 30 onto the conveyor 20 to which water is supplied by the water feeder 21. Thereby, the water on the conveyor 20 impregnates the hydraulic material 30 from the lower side.

  The hydraulic material 30 impregnated with water passes through the material spreading thickness adjusting roll 270 as the conveyor 20 advances, is pressed from above by the pressure roll 271, and is formed into a plate shape having a predetermined thickness.

  The hydraulic material 30 supplied onto the conveyor 20 by the hydraulic material supply machine 22 is a material mainly containing cement and containing an aggregate, a reinforcing material, a recycled material, or a combination thereof. .

  As the cement, for example, Portland cement, fly ash cement, or blast furnace cement is used. As the aggregate, for example, crushed stones such as granite and serpentine, quartzite powder, shirasu balloon, glass balloon, perlite, sand, or beads are used. Examples of the reinforcing material include cellulose-based pulp fibers, mineral fibers, organic resin fibers such as polypropylene and vinylon, glass fibers, and carbon fibers.

  The hydraulic material 30 is mixed with an admixture as necessary. As this admixture, for example, silica fume or siliceous raw materials such as powdered silica and fly ash are used.

  The hydraulic material supply machine 22 of the present embodiment includes a plurality of mixers 220 to 222 that mix and knead each raw material of the hydraulic material 30, and the homogeneity obtained by the plurality of mixers 220 to 222. A hydraulic material 30 is supplied onto the conveyor 20.

  The sprinkler 23 is provided above the conveyor 20 and sprays water onto the hydraulic material 30 on the conveyor 20 from above. Thereby, water is further added to the hydraulic material 30 that has passed through the material spreading thickness adjusting roll 270 and the pressure roll 271. The hydraulic material 30 to which water has been added by the sprinkler 23 is pressed from above by the pressure roll 272 by passing under the pressure roll 272 as the conveyor 20 advances, and has a plate shape having a predetermined thickness. To be molded.

  The cutting machine 24 is configured by a cutter roll disposed above the conveyor 20. The cutting machine 24 cuts the hydraulic material 30 on the conveyor 20 that has passed through the pressure roll 272 into a plurality of exterior material substrates 31.

  The cement material spreader 25 is disposed above the conveyor 20 and spreads the cement material 32 on the upper surface of the exterior material substrate 31 on the conveyor 20. The exterior material substrate 31 on which the cement material 32 is spread is pressed from above by the pressure roll 273 by passing under the pressure roll 273 and formed into a plate shape having a predetermined thickness.

  The colored stone spreader 26 is arranged above the conveyor 20 and spreads colored stone 33 on the cement material 32 on the exterior material substrate 31 that has passed through the pressure roll 273. The exterior material substrate 31 on which the colored stone 33 is spread by the colored stone spreader 26 is pressed from the upper side by the pressure roll 274 by passing below the pressure roll 274 and is formed into a plate shape having a predetermined thickness. Molded. In this way, a laminate 34 having the exterior material substrate 31, the cement material 32, and the chromite 33 is obtained.

  The laminated body 34 is cured after being inspected by an operator. And after painting, heating and drying, it is shipped as an exterior material.

(Production equipment abnormality diagnosis system)
The production facility abnormality diagnosis system monitors the state of the production facility 2 and diagnoses whether or not the state of the production facility 2 is abnormal by using the MT method.

  As shown in FIG. 2, the production facility abnormality diagnosis system includes an item group acquisition unit 10 and a diagnosis unit 11. The item group acquisition unit 10 acquires an item group including a plurality of items indicating the state of the production facility 2 at predetermined time intervals. The diagnosis unit 11 diagnoses whether the production facility 2 is abnormal using the MT method based on the item group acquired by the item group acquisition unit 10.

  The item group acquisition unit 10 includes a plurality of measuring devices 100 provided in the production facility 2. The measuring device 100 is provided in each of, for example, the conveyor 20, the water feeder 21, the hydraulic material feeder 22, the cutting machine 24, the cement material spreader 25, the chromite spreader 26, and the press machine 27, and the state of each device Measure items that serve as indicators.

  Items measured by the measuring device 100 include, for example, the speed of the belt conveyor 20, the amount of water supplied by the water supply device 21, the time and amount of pulverization of the raw material by the hydraulic material supply device 22, and the press pressure of the press machine 27. . The data of the item group acquired by the item group acquiring unit 10 is stored in the diagnosis unit 11 as needed.

  The diagnosis unit 11 is constituted by a computer including a CPU and a storage device, for example. The diagnosis unit 11 includes unit space creation means 110, Mahalanobis distance calculation means 111, and determination means 112.

  The unit space creating unit 110 creates a unit space (reference space) based on the item group acquired by the item group acquiring unit 10. The unit space is a collection of data serving as a reference when determining whether or not the state of the production facility 2 is abnormal.

  The Mahalanobis distance calculating unit 111 calculates the Mahalanobis distance MD based on the item group acquired by the item group acquiring unit 10 and the unit space created by the unit space creating unit 110.

  The determination unit 112 determines that the state of the production facility 2 is abnormal when the Mahalanobis distance MD calculated by the Mahalanobis distance calculation unit 111 exceeds a preset threshold value MDa, and when the state does not exceed the threshold value MDa. It is determined that the state of the production facility 2 is normal. Here, the threshold value MDa is set to a value smaller than a value when an abnormality occurs in the shape, dimensions, etc. of the exterior material actually produced by the production facility 2, for example.

(Diagnosis method)
The production facility abnormality diagnosis system diagnoses whether there is an abnormality in the production facility 2 as shown in FIG. 3 while the production facility 2 is operating.

  First, unit space creation data is acquired by the item group acquisition means 10 (step S1). This unit space creation data is data composed of a plurality of items, and the item group acquisition means 10 acquires an item group consisting of a plurality of items indicating the state of the production facility 2 at predetermined time intervals. can get. In this embodiment, the item group is acquired every minute, and the data of the item group for 60 consecutive minutes is stored as unit space creation data.

  Next, the unit space creating means 110 creates a unit space based on the item group data for creating the unit space described above (step S2).

  The unit space is created by a known method shown below. First, the unit space creation data is represented by an m × n matrix X. Here, m is the number of item groups, and n is the number of items in one group of item groups.

Next, the matrix X is normalized by (Equation 1) shown below to obtain a matrix X ′.
X ′ _ij = (X _ij −μ _j ) / σ _j (Equation 1)
Here, μ _j is m average values for each item, and σ _j is m standard deviations for each item. Next, a correlation matrix R between items is calculated from X ′ _ij . The correlation matrix R is an n × n square matrix. Then, an inverse matrix R ⁻¹ of the correlation matrix R, which is a unit space, is calculated from the correlation matrix R.

  After creating the unit space in step S2, the diagnosis unit 11 starts counting (step S3), acquires the item group by the item group acquisition means 10, and stores it (step S4).

  After step S4, the Mahalanobis distance calculation unit 111 calculates the Mahalanobis distance MD based on the unit space created in step S2 and the determination data including the item group acquired in step S4 (step S5). This Mahalanobis distance MD is calculated by a well-known method as shown below.

First, the determination data is represented by a vector y in the following (Expression 2).
y = [y ₁ y ₂ ... y _n ] (Expression 2)
Next, y is normalized by (Equation 3) shown below to obtain a vector Y.

Y _j = (y _j −μ _j ) / σ _j (Expression 3)
The vector Y is represented by the following (Formula 4).

Y = [Y ₁ Y ₂ ... Y _n ] (Formula 4)
The Mahalanobis distance MD is calculated using the following (Formula 5).

MD = Y · R ⁻¹ · Y ^T / n (Formula 5)
After calculating the Mahalanobis distance MD in step S5, the determination unit 112 compares the calculated Mahalanobis distance MD with a preset threshold value MDa to determine whether the Mahalanobis distance MD is equal to or less than the threshold value MDa ( Step S6).

  If the Mahalanobis distance MD is equal to or smaller than the threshold value MDa in step S6, the process proceeds to step S8. In step S6, if the Mahalanobis distance MD is larger than the threshold value MDa, the diagnosis unit 11 determines that the production facility 2 is abnormal, and notifies information indicating abnormality of the production facility 2 (step S7). The process proceeds to step S8. The notification is performed by, for example, display on a display device or sound generation by a speaker device.

  In step S8, the diagnosis unit 11 determines whether or not the predetermined time t1 has elapsed from the start of acquisition of the determination data based on the count started in step S3. The predetermined time t1 is a time for calculating the Mahalanobis distance MD using the same unit space, and the predetermined time t1 in the present embodiment is 30 minutes.

  If the predetermined time t1 has not elapsed in step S8, the process returns to step S4, and the item group (determination data) is acquired by the item group acquisition means 10 again. Thereby, acquisition of the item group by step S4 is performed continuously for every predetermined time t2. The predetermined time t2 of this embodiment is 1 minute.

  In step S8, if the predetermined time t1 has elapsed, the diagnosis unit 11 resets the count started in step S3 (step S9). Thereafter, the process proceeds to step S10, where the unit space creating means 110 newly sets the count. Start creating a simple unit space.

  FIG. 4 is an explanatory diagram showing a diagnosis method of the production facility diagnosis system of the present embodiment. In FIG. 4, a period surrounded by a rectangle is an item group acquisition period used to create each unit space. In FIG. 4, a period indicated by an arrow is a period in which an abnormality is diagnosed based on the corresponding unit space.

  In step S8, the unit space is created when the predetermined time t1 has elapsed, as shown in FIG. 4, at the point P when the unit space is created (the predetermined time t1 has elapsed since the start of the determination data acquisition). The item group acquisition means 10 in a time zone TZ between a first reference time point P1 that goes back from the first predetermined time T1 and a second reference time point P2 that goes back from the first reference time point P2 to the second predetermined time T2. This is done using the item group acquired by

  However, when the abnormality of the production facility 2 has been determined in the time zone TZ (that is, when the Mahalanobis distance MD has exceeded the threshold value MDa), the item group at this time is the unit space. The unit space is created using only the data of the item group when it is determined that the production facility 2 is normal, without being used as data for creating the.

  Specifically, in step S10 shown in FIG. 3, it is determined whether or not it may be determined that the production facility 2 is abnormal in the time zone TZ. In the present embodiment, the first reference time T1 and the second reference time T2 are each set to 1 hour. That is, the production facility 2 may be determined to be abnormal in one hour from the time point 2 hours before the time point P for creating the unit space to the time point 1 hour before the time point P for creating the unit space. It is determined whether or not. In the initial stage of diagnosis by the production facility diagnosis system, there may be a time zone TZ in which no item group is acquired in the time zone TZ. In this case, the subsequent calculation of the Mahalanobis distance MD is, for example, the first unit. Use space.

  In step S10, if it is not determined that the production facility 2 is abnormal in the time zone TZ, the unit space creation means 110 uses all the item groups acquired by the item group acquisition means 10 in the time zone TZ. A space is created (step 12).

  Further, if it is determined in step S10 that the production facility 2 is abnormal in the time zone TZ, the unit space creating unit 110 may select from a plurality of item groups acquired by the item group acquiring unit 10 in the time zone TZ. All the item groups when it is determined that the production facility 2 is abnormal are deleted (step S11), and then a unit space is created (step S12). That is, in this case, the unit space is created based on all the item groups except the item group in which the production facility 2 is determined to be abnormal among the item groups acquired in the time zone TZ.

  After step S12, the process returns to step S3. Thereafter, in the same manner as described above, the diagnosis unit 11 repeatedly calculates the Mahalanobis distance MD based on the unit space obtained in the most recent step S12, determines whether there is an abnormality in the production facility 2, and controls the creation of the unit space.

  As described above, the production facility abnormality diagnosis system according to the present embodiment calculates the Mahalanobis distance MD based on an item group including a plurality of items indicating the state of the production facility 2, and compares the Mahalanobis distance MD with the threshold value MDa. Since the presence or absence of abnormality of the production facility 2 is determined, the diagnosis of the production facility 2 can be easily performed even when there are many items indicating the state of the production facility 2.

  Moreover, in this embodiment, since the unit space used for calculation of the Mahalanobis distance MD is updated every predetermined time t1, the presence or absence of abnormality in the production facility 2 is diagnosed. Even when the state fluctuates, it is possible to accurately determine whether the production facility 2 is abnormal.

  In addition, the unit space in the present embodiment is created by the first reference time point P1 going back from the time point P for creating the unit space for the first predetermined time T1 and the second reference time point T2 going back from the first reference time point P1. In the time zone TZ between two reference time points P2, it is performed using the item group acquired by the item group acquisition means 10, and acquired in the time zone from the time point P for creating the unit space to the first reference time point P1. The item group is not used as data for creating a unit space. For this reason, even when the state of the production facility 2 gradually shifts to an abnormal state, it is possible to accurately determine whether the production facility 2 is abnormal.

  In this embodiment, the item group acquired when it is determined that the production facility 2 is abnormal is not used as data for creating a unit space. For this reason, the unit space can be created using only the data of the item group when the production facility 2 is normal, and thereby the presence or absence of abnormality of the production facility 2 can be accurately determined.

(Supplement)
The production facility abnormality diagnosis system of the present embodiment can be appropriately changed in design.

  For example, in this embodiment, the first unit space is created based on the item group acquired by the item group acquisition unit 10 in the time zone TZ immediately before the Mahalanobis distance MD is calculated. It may be set based on the data.

  In addition, when creating the unit space in steps S10 to S12, if it is determined that the production facility 2 is abnormal in the time zone TZ, the abnormality of the production facility 2 has not been determined in the past than the time zone TZ. The unit space created based on the item group acquired in the time zone TZ may be used for the subsequent calculation of the Mahalanobis distance MD. Also in this case, the unit space is created based on the item group excluding the item group determined to be abnormal by the determination unit 112.

  Further, when creating the unit space in steps S10 to S12, the item group acquired in all the time zones from the time point P when the unit space is created to a time point that is a predetermined time later is used as the data for creating the unit space. It may be.

  That is, the unit space may be created as the time zone TZ for creating the unit space based on the time point P at which the unit space is created and the item group from the time point P to a time point that is a predetermined time backward. That is, in this embodiment, T1 may be set to zero.

  Further, the production facility 2 is not limited to a facility that produces exterior materials, and the production facility abnormality diagnosis system can be applied to other production facilities 2 as long as the production facility 2 has a plurality of items indicating the state of the production facility 2. It is.

(effect)
The production facility abnormality diagnosis system of the embodiment described above is a production facility abnormality diagnosis system that diagnoses an abnormality in the production facility 2 using the MT method, and has the following characteristics. The production facility abnormality diagnosis system includes an item group acquisition unit 10, a unit space creation unit 110, a Mahalanobis distance calculation unit 111, and a determination unit 112. The item group acquisition unit 10 acquires an item group including a plurality of items indicating the state of the production facility 2 every predetermined time t2. The unit space creating unit 110 creates a unit space based on the item group acquired by the item group acquiring unit 10. The Mahalanobis distance calculating unit 111 calculates the Mahalanobis distance MD based on the item group acquired by the item group acquiring unit 10 and the unit space created by the unit space creating unit 110. The determination unit 112 determines that the state of the production facility 2 is abnormal when the Mahalanobis distance MD calculated by the Mahalanobis distance calculation unit 111 exceeds a preset threshold value MDa. The unit space creating unit 110 creates a unit space based on the item group excluding the item group determined to be abnormal by the determining unit 112 among the item groups acquired by the item group acquiring unit 10. Hereinafter, the production facility abnormality diagnosis system having this feature is referred to as the production facility abnormality diagnosis system according to the first aspect.

  In the production facility abnormality diagnosis system of the first aspect, the unit space is determined based on the item group excluding the item group determined to be abnormal by the determination unit 112 among the item groups acquired by the item group acquisition unit 10. create. For this reason, it is possible to create a unit space using only the data of the item group when it is determined that the production facility 2 is normal, thereby accurately determining whether the production facility 2 is abnormal.

  Further, the production facility abnormality diagnosis system of the above embodiment has the following additional features in addition to the features of the production facility abnormality diagnosis system of the first aspect. The unit space creating means 110 is between the first reference time P1 that goes back for the first predetermined time T1 from the time P when the unit space is created and the second reference time P2 that goes back for the second predetermined time T2 from the first reference time P1. The unit space is created based on the item group excluding the item group determined to be abnormal by the determination unit 112 among the item groups acquired by the item group acquisition unit 10. Hereinafter, the production facility abnormality diagnosis system having this feature is referred to as a production facility abnormality diagnosis system according to the second aspect.

  In the production facility abnormality diagnosis system of the second aspect, the item group acquired in the time zone from the time point P at which the unit space is created to the first reference time point P1 is not used as data for creating the unit space. For this reason, even when the state of the production facility 2 gradually shifts to an abnormal state, it is possible to accurately determine whether the production facility 2 is abnormal.

MD Mahalanobis distance MDa threshold P Time point when unit space is created P1 First reference point P2 Second reference point T1 First predetermined time T2 Second predetermined time t2 Predetermined time 10 Item group acquisition means 110 Unit space generation means 111 Mahalanobis distance calculation means 112 Judgment means 2 Production equipment

Claims (2)

A production facility abnormality diagnosis system for diagnosing production facility abnormality using the MT method,
Item group acquisition means for acquiring an item group consisting of a plurality of items indicating the state of the production facility every predetermined time;
Unit space creating means for creating a unit space based on the item group acquired by the item group acquiring means;
Mahalanobis distance calculating means for calculating a Mahalanobis distance based on the item group acquired by the item group acquiring means and the unit space created by the unit space creating means;
Determining means for determining that the state of the production facility is abnormal when the Mahalanobis distance calculated by the Mahalanobis distance calculating means exceeds a preset threshold;
The unit space creating means creates the unit space based on the item group excluding the item group determined to be abnormal by the determining means among the item group acquired by the item group acquiring means. Characteristic production equipment abnormality diagnosis system.   The unit space creating means includes the item group between a first reference time that is a first predetermined time back from the time when the unit space is created and a second reference time that is a second predetermined time back from the first reference time. 2. The production facility abnormality according to claim 1, wherein the unit space is created based on the item group excluding the item group determined to be abnormal by the determination unit among the item group acquired by the acquisition unit. Diagnostic system.

Images