JP2022037394A - Abnormality detection device, sluice system, and abnormality detection method - Google Patents

Abstract

To properly detect abnormality in a radial gate.SOLUTION: In a radial gate, a door body and a driving portion are connected by a connection medium. In the door body, a skin plate and a bearing portion are connected by a leg post. A load value acquisition portion 46 of an abnormality detection device 40 acquires a value representing a load applied to the connection medium as a bearing load value. An evaluation value acquisition portion 47 acquires a value representing a stress applied to the leg post as a leg post evaluation value. A normal value estimation portion 41 estimates a normal value of the leg post evaluation value based on a water level at an installation position of the radial gate, an opening of the door body, and the bearing load value. A determination portion 42 determines presence/absence of abnormality in the radial gate by comparing the leg post evaluation value acquired by the evaluation value acquisition portion 47 with the normal value estimated by the normal value estimation portion 41. Thus, abnormality in the radial gate can be properly detected.SELECTED DRAWING: Figure 3

Description

The present invention relates to an anomaly detection device, a floodgate system and an anomaly detection method.

Conventionally, radial gates have been used in dams and the like. The manager of the radial gate strives to confirm and maintain its soundness by conducting manual inspections and surveys (monthly inspections, annual inspections, detailed surveys once every ten years, etc.).

In addition, Non-Patent Document 1 discloses a method of monitoring the strain of the pedestal in the radial gate and estimating the coefficient of friction in the bearing portion. Specifically, the analysis value of the strain change amount of the pedestal is obtained by finite element analysis with the acting water level on the gate and the bearing friction coefficient as parameters, and the bearing friction coefficient is obtained from the actual working water level and the strain change amount. Presumed.

Yuzo Shiogama, "Friction Resistance Monitoring of Radial Gate Bearings", JSCE Proceedings A Vol. 65 No. l, January 2009, pp. 1-14

By the way, when the radial gate is aged, an abnormality in the bearing part (rotating part) may cause the door body to rotate poorly or the stress of the pedestal to be excessive. The state of the part is not usually grasped. Further, in the above-mentioned inspection and investigation, only the operation data for a limited period can be obtained, and the tendency management in the case where the aging state progresses rapidly cannot be performed. Further, since the operating conditions for inspection (state without water pressure, etc.) are different from the conditions during normal operation, it may not be possible to grasp the normal conditions.

In Non-Patent Document 1, the friction coefficient at the bearing portion is estimated, but it is not determined whether or not the estimated value of the friction coefficient is normal. In addition, since modeling for finite element analysis is performed with design dimensions, it is not possible to obtain results that take into account the effects of installation errors, manufacturing errors, and the like, that is, the reproducibility of field conditions is low. Further, since the friction coefficient is estimated at one point at the start of operation of the radial gate, it is not possible to continuously grasp the state during operation.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a novel method capable of appropriately detecting an abnormality in a radial gate.

The invention according to claim 1 is an abnormality detecting device for detecting an abnormality in a radial gate, in which a door body and a driving unit are connected by a connecting medium in the radial gate, and a value indicating a load applied to the connecting medium. The load value acquisition unit that acquires the support load value, and the skin plate and the support portion of the door body are connected by a pedestal, and the value indicating the stress applied to the pedestal is acquired as the pedestal evaluation value. The value acquisition unit, the normal value estimation unit that estimates the normal value of the pedestal evaluation value based on the water level at the installation position of the radial gate, the opening degree of the door body, and the support load value, and the evaluation value. A determination unit for determining the presence or absence of an abnormality in the radial gate is provided by comparing the pedestal evaluation value acquired by the acquisition unit with the normal value estimated by the normal value estimation unit.

The invention according to claim 2 is the abnormality detection device according to claim 1, wherein each side portion of the skin plate comes into contact with a door stop via a watertight rubber, and the normal value estimation unit applies water pressure. Based on the length of the watertight rubber received, the normal value is estimated when the door body rotates.

The invention according to claim 3 is the abnormality detection device according to claim 1 or 2, wherein the normal value estimation unit is based on the flow rate of water passing through the radial gate when the door body is rotated. The normal value is estimated.

The invention according to claim 4 is the abnormality detection device according to any one of claims 1 to 3, wherein the normal value estimation unit starts rotation of the door body when the door body rotates. The normal value is estimated based on the elapsed time from.

The invention according to claim 5 is the abnormality detection device according to any one of claims 1 to 4, further comprising an inclinometer attached to the pedestal.

The invention according to claim 6 is the abnormality detection device according to any one of claims 1 to 5, and the pedestal evaluation acquired by the evaluation value acquisition unit at one time point in a normal state. It is generated by learning using a plurality of teacher data at a plurality of time points, using a combination of the value and the feature amount group substantially including the water level, the opening degree and the supporting load value at the one time point as the teacher data. Using the trained model, the normal value estimation unit estimates the normal value of the pedestal evaluation value.

The invention according to claim 7 is an abnormality detecting device for detecting an abnormality in a radial gate, in which a door body and a driving unit are connected by a connecting medium in the radial gate, and a value indicating a load applied to the connecting medium. A load value acquisition unit that acquires the support load value as a support load value, a normal value estimation unit that estimates a normal value of the support load value based on the water level at the installation position of the radial gate, and the opening degree of the door body, and the above. A determination unit for determining the presence or absence of an abnormality in the radial gate is provided by comparing the support load value acquired by the load value acquisition unit with the normal value estimated by the normal value estimation unit.

The invention according to claim 8 is a floodgate system, comprising a radial gate and the abnormality detection device according to any one of claims 1 to 7, which detects an abnormality in the radial gate.

The invention according to claim 9 is an abnormality detecting method for detecting an abnormality in a radial gate, wherein a) a door body and a driving unit are connected by a connecting medium in the radial gate, and a load applied to the connecting medium is applied. The step of acquiring the indicated value as a supporting load value, and b) the skin plate and the support portion are connected by a pedestal in the door body, and the value indicating the stress applied to the pedestal is acquired as the pedestal evaluation value. Steps and c) a step of estimating a normal value of the pedestal evaluation value based on the water level at the installation position of the radial gate, the opening degree of the door body, and the supporting load value, and d) the step b). The step includes a step of determining the presence or absence of an abnormality in the radial gate by comparing the pedestal evaluation value acquired in the above step with the normal value estimated in the step c).

The invention according to claim 10 is an abnormality detecting method for detecting an abnormality in a radial gate, wherein a) a door body and a driving unit are connected by a connecting medium in the radial gate, and a load applied to the connecting medium is applied. A step of acquiring the indicated value as a support load value, b) a step of estimating a normal value of the support load value based on the water level at the installation position of the radial gate, and the opening degree of the door body, and c) the step of estimating the normal value of the support load value. It includes a step of determining the presence or absence of an abnormality in the radial gate by comparing the supporting load value acquired in the step a) with the normal value estimated in the step b).

According to the present invention, anomalies in the radial gate can be appropriately detected.

It is a figure which shows the structure of the floodgate system which concerns on 1st Embodiment. It is a figure which shows the rotating door body. It is a block diagram which shows the functional structure of an abnormality detection apparatus. It is a figure which shows the flow of the process which detects the abnormality of a radial gate. It is a figure which shows the door body to which the inclinometer is attached. It is a block diagram which shows the functional structure of the abnormality detection apparatus which concerns on 2nd Embodiment. It is a figure which shows the flow of the process which detects the abnormality of a radial gate.

FIG. 1 is a diagram showing a configuration of a floodgate system 1 according to a first embodiment of the present invention. The floodgate system 1 includes a radial gate 10 which is a floodgate, and a computer 4. The radial gate 10 of FIG. 1 is installed in a dam, and the computer 4 is installed, for example, in a building adjacent to the dam. The radial gate 10 may be provided in equipment other than the dam (for example, a dike). Further, the computer 4 may be provided at a position far away from the radial gate 10. In this case, various measuring units attached to the radial gate 10 are connected to a communication device or the like, and can communicate with the computer 4 via a network such as the Internet. The computer 4 may be a computer provided by a cloud service.

The radial gate 10 includes a door body 2 and a switchgear 3. The door body 2 is installed in the opening 91 provided in the embankment body of the dam, and is used to regulate the flow of water in the opening 91. The door body 2 is formed of, for example, metal. The door body 2 includes a skin plate 21 and two plate support portions 22. The skin plate 21 is a plate-shaped member substantially parallel to the rotation axis J1 in FIG. The cross-sectional shape of the skin plate 21 perpendicular to the rotation axis J1 is a substantially arc shape having a predetermined radius about the rotation axis J1. The skin plate 21 is provided between two side surfaces 911 (hereinafter referred to as "opening side surface 911") of the opening 91 substantially perpendicular to the rotation axis J1.

The two plate support portions 22 are provided apart from each other in the direction of the rotation axis J1 and are typically arranged in the vicinity of the two opening side surfaces 911. FIG. 1 shows only one plate support portion 22. Each plate support portion 22 includes a support portion 23 and a plurality of pedestals 24. The support portion 23 is a rotating portion of the door body 2, and includes, for example, a trunnion hub 231 and a trunnion pin 232. The trunnion pin 232 is fixed to the dam body (eg, opening side surface 911). The trunnion hub 231 is a cylindrical member and is fitted into the trunnion pin 232 via a radial bearing. The trunnion hub 231 is rotatably supported around the rotation axis J1 by the trunnion pin 232. The axis of rotation J1 coincides with the center of the trunnion pin 232. In the bearing portion 23, a thrust bearing may be further provided with respect to the trunnion hub 231. Depending on the design of the bearing 23, the trunnion hub 231 may be fixed to the embankment and the trunnion pin 232 may be rotatably supported by the trunnion hub 231.

The plurality of pedestals 24 are long members extending from the rotating body of the support portion 23 (here, the trunnion hub 231) toward the skin plate 21. Typically, the plate support 22 has two pedestals 24. The two pedestals 24 form an acute angle and are fixed to the rotating body of the support portion 23. Each pedestal 24 is, for example, H-shaped steel, I-shaped steel, or the like. Two girders 25 are provided on the surface of the skin plate 21 on the side of the support portion 23. Each girder 25 is a long member substantially parallel to the rotation axis J1. In the two pedestals 24, the ends opposite to the support portion 23 are fixed to the two girders 25, respectively. In the door body 2, the skin plate 21 and the support portion 23 are connected by a girder 25 and a pedestal 24.

A pulley (sheave) 26 is further provided on each plate support portion 22. The pulley 26 is a moving pulley that moves (rotates) together with the door body 2. In the example of FIG. 1, the pulley 26 is arranged in the vicinity of the lower girder 25. The number and position of the pulleys 26 may be changed as appropriate. In the door body 2, a reinforcing member may be appropriately provided between the two pedestals 24 or on the surface of the skin plate 21 on the support portion 23 side.

The side portion 211 of the skin plate 21 facing each opening side surface 911 has a surface substantially parallel to the opening side surface 911. The side portion 211 is provided with a watertight rubber 212 (hereinafter referred to as "side watertight rubber 212"). The side watertight rubber 212 is provided over substantially the entire area from the upper end to the lower end of the side portion 211 in FIG. Similar to the cross-sectional shape of the skin plate 21 perpendicular to the rotation axis J1, the side watertight rubber 212 has a substantially arc shape with a predetermined radius centered on the rotation axis J1. The side portion 211 of the skin plate 21 comes into contact with the opening side surface 911, which is a door stop, via the side watertight rubber 212. The side watertight rubber 212 is in close contact with the opening side surface 911 to prevent water from passing between the side portion 211 and the opening side surface 911. It is preferable that each side portion 211 of the skin plate 21 is provided with a side roller (not shown) that rolls on the opening side surface 911. As a result, the door body 2 can smoothly rotate about the rotation axis J1.

In the skin plate 21, a watertight rubber 214 (hereinafter referred to as “lower watertight rubber 214”) is also provided at the lower end portion 213. The lower watertight rubber 214 is provided at the lower end portion 213 from the vicinity of one opening side surface 911 to the vicinity of the other opening side surface 911, that is, substantially the entire lower end portion 213. The lower watertight rubber 214 extends substantially parallel to the axis of rotation J1. The lower end portion 213 of the skin plate 21 comes into contact with the bottom surface 912 of the opening 91, which is a door stop, via the lower watertight rubber 214. When the lower watertight rubber 214 is in close contact with the bottom surface 912, water is prevented from passing between the lower end portion 213 and the bottom surface 912. When the radial gate 10 is provided in the middle of the dam, a watertight rubber is also provided at the upper end of the skin plate 21.

The opening / closing device 3 is, for example, a wire rope winch type, and the door body 2 is rotated to open / close the door body 2 (opening / closing the opening 91). The switchgear 3 includes, for example, a drive unit 31 and two metal wire ropes 36. The drive unit 31 has two drums 311. The two drums 311 are respectively located above the two opening side surfaces 911. As will be described later, a part of the wire rope 36 is wound around each drum 311. FIG. 1 illustrates only one drum 311 and one wire rope 36. The number of drums 311 and wire rope 36 may be changed arbitrarily.

The two drums 311 are connected to a shaft (not shown) extending in the direction of the rotation shaft J1 via a gear or the like. The shaft is connected to the motor via a speed reducer or the like. In the drive unit 31, the electric motor rotates the shaft, so that the two drums 311 rotate in the same rotation direction. In the drive unit 31, the rotation direction of the drum 311 can be switched between forward rotation and reverse rotation.

One end of the wire rope 36, which is a connecting medium, is fixed to each drum 311 and a part of the wire rope 36 is wound around the drum 311. The portion of each wire rope 36 that is not wound around the drum 311 is hung on the pulley 26 of the plate support portion 22. The other end of the wire rope 36 is held by the load value acquisition unit 46 described later in the vicinity of the drum 311. The load value acquisition unit 46 is fixed to the embankment body. In the radial gate 10, the door body 2 and the drive unit 31 are connected by a wire rope 36. In the switchgear 3, a rope other than the wire rope 36 or another rope-like member such as a chain may be used as the connecting medium.

In the switchgear 3, when the drum 311 rotates in the normal direction, the wire rope 36 is wound around the drum 311 and the door body 2 in FIG. 1 rotates clockwise around the rotation axis J1. As a result, as shown in FIG. 2, the skin plate 21 moves (that is, rises) in a direction away from the bottom surface 912 of the opening 91. Further, when the drum 311 is reversed, the wire rope 36 is sent out from the drum 311 and the door body 2 rotates counterclockwise about the rotation axis J1. As a result, the skin plate 21 moves (that is, descends) toward the bottom surface 912 of the opening 91. In this way, the door body 2 is opened and closed at the opening 91.

The radial gate 10 of FIG. 1 further includes a load value acquisition unit 46, an evaluation value acquisition unit 47, a water level meter 11, and an opening degree meter 12. The load value acquisition unit 46 and the evaluation value acquisition unit 47 are configured to be included in the abnormality detection device 40 (see FIG. 3) described later. As described above, the load value acquisition unit 46 is connected to the wire rope 36. The load value acquisition unit 46 has, for example, a load cell or the like, and measures a load (here, tension) applied to the wire rope 36 as a support load value. The load value acquisition unit 46 may measure the load by using a device other than the load cell. In the example of FIG. 1, the load value acquisition portion 46 is connected to the wire rope 36 hung on the pulley 26 of each plate support portion 22. The load value acquisition unit 46 is individually provided for each plate support unit 22.

The evaluation value acquisition unit 47 is attached to a predetermined position (hereinafter, referred to as “evaluation position”) of the pedestal 24 of each plate support portion 22, and acquires a value indicating stress applied to the pedestal 24 as a pedestal evaluation value. .. In the example of FIG. 1, the evaluation value acquisition unit 47 is a strain gauge, and acquires the strain of the pedestal 24 as the pedestal evaluation value. The evaluation value acquisition unit 47 may be other than a strain gauge as long as it can acquire a value substantially indicating the stress applied to the pedestal 24. For example, the evaluation value acquisition unit 47 has a configuration in which the strain and the amount of deflection of the pedestal 24 are measured using a length measuring device, a camera, or the like, and a configuration in which the stress acting on the pedestal 24 is measured using an infrared camera. May be used as. The evaluation value acquisition unit 47 may be attached to a plurality of pedestals 24, or may be attached to only one pedestal 24. In the present embodiment, the load value acquisition unit 46 and the evaluation value acquisition unit 47 are provided for each of the two plate support units 22, but may be provided for only one plate support unit 22.

The water level gauge 11 is, for example, a float type water level gauge, and measures the water level at the installation position of the radial gate 10 (that is, the water level of water in contact with the skin plate 21) by detecting the position of the float floating on the water surface. do. In this embodiment, the water level is the height from the bottom of the dam to the water surface. The water level gauge 11 may be a sound wave type, ultrasonic type, pressure type water level gauge, or the like that measures the water level without using a float. The opening meter 12 detects the opening degree of the door body 2 by using, for example, a wire separately attached to the door body 2. In the examples of FIGS. 1 and 2, the opening degree of the door body 2 is the height from the bottom surface 912 of the opening 91 (to be exact, the contact position of the lower watertight rubber 214) to the lower end of the skin plate 21 (in the vertical direction). Distance). The opening degree meter 12 may detect the opening degree of the door body 2 by using various sensors or the like.

FIG. 3 is a block diagram showing a functional configuration of the abnormality detection device 40. The abnormality detection device 40 detects an abnormality in the radial gate 10, and includes a load value acquisition unit 46, an evaluation value acquisition unit 47, a normal value estimation unit 41, and a determination unit 42. As described above, the load value acquisition unit 46 is connected to each wire rope 36 in FIG. 1 and acquires a support load value indicating the load applied to the wire rope 36. Further, the evaluation value acquisition unit 47 is attached to a predetermined evaluation position of the pedestal 24, and acquires the pedestal evaluation value indicating the stress applied to the pedestal 24.

The normal value estimation unit 41 in FIG. 3 estimates the normal value of the pedestal evaluation value. The normal value of the pedestal evaluation value is the value of the pedestal evaluation value that should occur when it is assumed that no abnormality has occurred in the radial gate 10 under the current operating conditions. The determination unit 42 determines whether or not there is an abnormality in the radial gate 10. The functions of the normal value estimation unit 41 and the determination unit 42 are realized by the CPU of the computer 4 or the like executing arithmetic processing according to a predetermined program (that is, by executing the program by the computer). The functions of the normal value estimation unit 41 and the determination unit 42 may be realized by a dedicated electric circuit, or a partially dedicated electric circuit may be used. In FIG. 3, the water level gauge 11 and the opening gauge 12 are also shown by blocks.

FIG. 4 is a diagram showing a flow of processing in which the abnormality detection device 40 detects an abnormality in the radial gate 10. In the floodgate system 1 shown in FIG. 1, for example, when the door body 2 is rotated (moved), that is, when the skin plate 21 is raised or lowered, the load value acquisition unit 46 acquires the support load value in the wire rope 36. (Step S11). Further, the evaluation value acquisition unit 47 acquires the pedestal evaluation value in the pedestal 24 (step S12). As will be described later, the support load value and the pedestal evaluation value are repeatedly measured during the period from the start to the end of the rotation of the door body 2 (for example, several minutes). The support load value is input to the normal value estimation unit 41, and the measured value of the pedestal evaluation value is input to the determination unit 42. In the following description, attention will be paid to the measured values (supporting load value and pedestal evaluation value) acquired by the load value acquisition unit 46 and the evaluation value acquisition unit 47 of one plate support unit 22, while the other plate support unit 22 The same processing is performed for the measured values acquired by the load value acquisition unit 46 and the evaluation value acquisition unit 47.

Further, in the radial gate 10, the water level is acquired by the water level meter 11, and the opening degree of the door body 2 is acquired by the opening meter 12. The water level and the opening degree of the door body 2 are input to the normal value estimation unit 41. In the normal value estimation unit 41, the length of the portion of the side watertight rubber 212 provided on each side portion 211 of the skin plate 21 based on the water level and the opening degree of the door body 2 (hereinafter referred to as “side”). The length of the rubber submergence is required.) As described above, the side watertight rubber 212 has a substantially arc shape with a predetermined radius centered on the rotation axis J1. The length of the side rubber submergence in the side watertight rubber 212 of FIG. 2 is, for example, the height of the lower end of the skin plate 21 specified by the opening degree of the door body 2 on the circumference of the radius centered on the rotation axis J1. Therefore, it can be obtained as the length of the arc up to the height of the water surface specified by the water level. In the side watertight rubber 212, the frictional force between the portion located in the water and the opening side surface 911 becomes larger than the frictional force between the portion located in the air and the opening side surface 911 due to the influence of water pressure. The side rubber submersion length is the length of the side watertight rubber 212 that receives water pressure.

In the normal value estimation unit 41, the water level acquired at about the same time, the opening degree of the door body 2, the supporting load value, and the side rubber submersion length are treated as a feature quantity group (feature vector). As will be described later, the feature amount group may include other values (feature amounts). In the abnormality detection device 40 of FIG. 3, the estimator 411, which is a trained model, is generated in advance by machine learning described later and introduced into the normal value estimation unit 41. In the estimator 411 of the normal value estimation unit 41, by inputting the feature amount group, the normal value of the pedestal evaluation value (here, strain) indicating the stress at the evaluation position of the pedestal 24 is output. That is, the normal value of the pedestal evaluation value is estimated based on the feature amount group (step S13). The normal value of the pedestal evaluation value is input to the determination unit 42.

Here, in the pedestal 24, the moment due to the tension of the wire rope 36, the moment due to the weight of the door body 2, the moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911, and the rotating body of the bearing portion 23. The moment due to the frictional force between the bearing (radial bearing and thrust bearing) acts mainly. At the evaluation position where the evaluation value acquisition unit 47 is attached to the pedestal 24, the pedestal evaluation value indicating stress also changes due to the change in the magnitude of these moments.

The moment due to the tension of the wire rope 36 correlates with the supporting load value. The moment due to the weight of the door body 2 is affected by the opening degree of the door body 2 (inclination angle of the pedestal 24). The moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911 is affected by the submersion length of the side rubber. The moment due to the frictional force between the rotating body of the bearing portion 23 and the bearing is affected by the water level (water pressure). Therefore, in the estimator 411 generated by the learning described later, the normal value of the pedestal evaluation value is obtained from the feature amount group including the water level, the opening degree of the door body 2, the side rubber submergence length, and the supporting load value. It can be estimated.

The determination unit 42 compares the measured value of the pedestal evaluation value acquired by the evaluation value acquisition unit 47 in step S12 with the normal value of the pedestal evaluation value estimated by the normal value estimation unit 41 in step S13. .. For example, when the absolute value of the difference between the measured value of the pedestal evaluation value and the normal value is larger than a predetermined threshold value, it is determined that some abnormality has occurred in the radial gate 10. In this case, for example, a display indicating that an abnormality has occurred is displayed on the display of the computer 4. Notification indicating the occurrence of an abnormality may be performed by turning on a light, sounding a buzzer, notifying an information communication terminal using e-mail or the like.

Further, when the absolute value of the difference is equal to or less than the threshold value, it is determined that no abnormality has occurred in the radial gate 10, and the notification indicating the occurrence of the abnormality is not performed. In this way, the determination unit 42 determines whether or not there is an abnormality in the radial gate 10 (step S14). The determination of the abnormality in the determination unit 42 may be performed based on a change in the distribution of the difference over a certain period of time, or any statistical method may be used. Further, the presence or absence of abnormality may be determined based on the ratio of the measured value of the pedestal evaluation value to the normal value and the like.

In the abnormality detection device 40, the processes of steps S11 to S14 are repeated from the start to the end of the rotation of the door body 2 (step S15). In the determination unit 42, the period until parts replacement or inspection is determined according to the magnitude of the absolute value of the difference between the measured value of the pedestal evaluation value and the normal value, and the period may be notified. .. In this way, the process of determining the period until parts replacement or inspection can be substantially regarded as the detection of abnormality due to aged deterioration (determination of the presence or absence of abnormality).

Next, the generation of the estimator 411 will be described. Here, the estimator (regressor) uses a decision tree, a neural network, or the like. Generating an estimator means generating an estimator by assigning values to parameters included in the estimator, determining the structure of the estimator, and the like. The generation of the estimator 411 may be performed by the computer 4 of the floodgate system 1 or by another external computer. In the generation of the estimator 411, any statistical method or machine learning method other than the above-mentioned method may be used.

In the water gate system 1, in the past operation, the support load value acquired by the load value acquisition unit 46 and the measured value of the pedestal evaluation value acquired by the evaluation value acquisition unit 47 are associated with the date and time. It is stored in the computer 4. The same applies to the water level measured by the water level gauge 11 and the opening degree of the door body 2 measured by the opening degree meter 12. Furthermore, the period during which the abnormality occurred in the radial gate 10 is also recorded.

In the generation of the estimator 411, a plurality of teacher data are prepared at a time point when no abnormality occurs, that is, at a plurality of time points which are normal times. Each teacher data is a combination of the measured value of the pedestal evaluation value by the evaluation value acquisition unit 47 at one time point in the normal state and the feature amount group at that time point. The feature amount group is the same as the feature amount group used in step S13, and includes the water level at that time, the opening degree of the door body 2, the side rubber submersion length, and the supporting load value.

When multiple teacher data are prepared, learning is performed so that the output of the estimator for the input of the feature quantity group in the multiple teacher data and the measured value of the pedestal evaluation value indicated by the multiple teacher data are almost the same. This is done and a trained model, estimator 411, is generated. The estimator 411 may be generated by a well-known method, and a value close to the measured value of the pedestal evaluation value indicated by the teacher data is output for the input of the feature amount group of each teacher data. In addition, the values of the parameters included in the estimator and the structure of the estimator are determined. The estimator 411 (actually, the value of the parameter and the information indicating the structure of the estimator 411) is transferred to the normal value estimation unit 41 and introduced. As described above, the normal value estimation unit 41 estimates the normal value of the pedestal evaluation value using the estimator 411.

As described above, the load value acquisition unit 46 of the abnormality detection device 40 acquires a support load value indicating the load applied to the wire rope 36. Further, in the normal value estimation unit 41, the water level, the opening degree of the door body 2, the side rubber submerged length, and the side rubber submerged length are obtained by using the water level and the opening degree of the door body 2. The normal value of the pedestal evaluation value is estimated based on the feature quantity group including the support load value. As described above, the moment due to the frictional force between the rotating body of the bearing portion 23 and the bearing is affected by the water level (water pressure). The moment due to the weight of the door body 2 is affected by the opening degree of the door body 2 (inclination angle of the pedestal 24). The moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911 is affected by the submersion length of the side rubber. The moment due to the tension of the wire rope 36 correlates with the supporting load value. The normal value estimation unit 41 can accurately estimate the normal value of the pedestal evaluation value based on the feature quantity group including the influence of on-site conditions such as installation error and manufacturing error.

Further, the evaluation value acquisition unit 47 acquires the measured value of the pedestal evaluation value indicating the stress applied to the pedestal 24. Then, in the determination unit 42, the measured value of the pedestal evaluation value acquired by the evaluation value acquisition unit 47 and the normal value of the pedestal evaluation value estimated by the normal value estimation unit 41 are compared, and the radial gate 10 is used. The presence or absence of abnormality is determined. As a result, it is possible to appropriately detect an abnormality in the radial gate 10, such as an aging (rotational failure) state of the bearing portion 23. Actually, it is possible to determine whether or not there is a margin for the destruction of the pedestal 24 based on the measured value of the pedestal evaluation value. In the floodgate system 1 including the radial gate 10 and the abnormality detection device 40, even if an abnormality occurs in the radial gate 10, it can be dealt with promptly, and stable operation can be realized.

In the abnormality detecting device 40, it is preferable that the presence or absence of an abnormality is repeatedly determined when the door body 2 is rotated. As a result, the state of the radial gate 10 can be continuously grasped during the rotation operation. By continuously comparing the measured value of the pedestal evaluation value acquired by the abnormality detection device 40 with the normal value, it is possible to manage the tendency of the radial gate 10 in an aging state.

By the way, as described above, the side rubber submersion length is obtained from the water level and the opening degree of the door body 2. Therefore, even if the side rubber submersion length is omitted from the feature quantity group, the pedestal evaluation value is performed with a certain degree of accuracy based on the water level, the opening degree of the door body 2, and the supporting load value. It is possible to estimate the normal value of. Similarly, the water level and the opening degree of the door body 2 may be omitted from the feature quantity group, and in this case as well, the legs are used with a certain degree of accuracy by using the side rubber submersion length and the supporting load value. It is possible to estimate the normal value of the column evaluation value. Since the side rubber submersion length is obtained from the water level and the opening degree of the door body 2, even when the feature amount group includes only the side rubber submergence length and the supporting load value, the water level is substantially the same. It can be said that the normal value of the pedestal evaluation value is estimated based on the opening degree of the door body 2 and the supporting load value.

By using the trained model (estimator 411) generated by learning, the normal value estimation unit 41 can accurately acquire the normal value of the pedestal evaluation value. As a result, it is possible to accurately detect the abnormality in the radial gate 10. In the generation of the trained model and the estimation of the normal value using the trained model, as described above, the feature quantity group including the water level, the opening degree of the door body 2, and the supporting load value, or the side rubber. A feature quantity group including a submerged length and a supporting load value may be used, and the side rubber submerged length is obtained from the water level and the opening degree of the door body 2. Therefore, it can be said that the feature amount group substantially includes the water level, the opening degree of the door body 2, and the supporting load value.

Depending on the design of the anomaly detection device 40, statistical regression analysis is performed using the above feature quantity group at normal times as an explanatory variable and the pedestal evaluation value (normal value) as an objective variable, and the obtained regression equation is used. Then, in step S13 of FIG. 4, the normal value of the pedestal evaluation value may be estimated from the feature amount group. The normal value of the pedestal evaluation value may be estimated using various statistical methods or machine learning methods.

The feature amount group may include a value (feature amount) other than the water level, the opening degree of the door body 2, the side rubber submersion length, and the supporting load value. For example, the direction of the moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911 depends on the rotation direction of the door body 2. Therefore, the feature amount group is a value indicating whether the rotation direction of the door body 2 in FIG. 1 is clockwise or counterclockwise (whether the movement direction of the skin plate 21 is ascending or descending). Is preferably included.

Further, in the preferable abnormality detection device 40, the flow rate of water passing through the radial gate 10 (hereinafter referred to as “discharge flow rate”) is included in the feature amount group. The discharge flow rate can be calculated using a predetermined calculation formula from the width of the opening 91 (width in the direction of the rotation axis J1), the water level, the opening of the door body 2, and the like. When the skin plate 21 rises or falls, an upward or downward force depending on the discharge acts at a specific opening degree depending on the shape of each door body. For example, when the door body 2 is rotated so that the skin plate 21 rises, a force corresponding to the discharge flow rate acts on the skin plate 21 due to the influence of water passing while contacting the lower end of the skin plate 21. Further, when the door body 2 is rotated so that the skin plate 21 descends, the water flowing below the skin plate 21 exerts a resistance force that hinders the descending of the skin plate 21. Therefore, the normal value estimation unit 41 depends on the discharge flow rate by estimating the normal value of the pedestal evaluation value at the time of rotation of the door body 2 based on the flow rate (discharge flow rate) of the water passing through the radial gate 10. Considering the load, the normal value of the pedestal evaluation value can be estimated more accurately. When the discharge flow rate is included in the feature amount group, it is preferable that the rotation direction of the door body 2 is also included in the feature amount group.

Further, in the radial gate 10, when the door body 2 stopped at a certain rotation position rotates from the rotation position in any rotation direction, the door body 2 is said to have the same other conditions. The measured value of the pedestal evaluation value obtained at the time of this rotation may differ depending on whether the rotation to the rotation position (the previous rotation) was clockwise or counterclockwise. In this way, one of the reasons why the measured value of the pedestal evaluation value at the time of this rotation depends on the rotation direction at the time of the previous rotation is that when the previous rotation is completed, the bearing portion 23 becomes a rotating body. It is considered that the frictional force generated between the bearing and the bearing remains until the time of this rotation.

Therefore, in the preferred abnormality detecting device 40, when the door body 2 is rotated, the rotation direction of the door body 2 this time and the rotation direction of the previous rotation are included in the feature quantity group. That is, the normal value estimation unit 41 estimates the normal value of the pedestal evaluation value based on the rotation direction of the door body 2 this time and the rotation direction of the previous rotation when the door body 2 is rotated. do. As a result, the normal value of the pedestal evaluation value can be estimated more accurately in consideration of the frictional force remaining on the bearing portion 23 and the like.

Further, the elapsed time from the start of rotation of the door body 2 may be included in the feature quantity group. Depending on the structure of the door body 2, water may be taken into a part of the door body 2 (for example, near the pedestal 24 side of the skin plate 21). In this case, the elapsed time from the start of rotation of the door body 2 As a result, the amount of water taken into the door body 2 changes. Further, the contact position between the rotating body of the bearing portion 23 and the bearing, the contact position between the side portion 211 of the skin plate 21 and the opening side surface 911, and the like also change depending on the elapsed time. Therefore, the normal value estimation unit 41 estimates the normal value of the pedestal evaluation value based on the elapsed time from the start of rotation of the door body 2 when the door body 2 is rotating, so that the operating load fluctuates during operation. The normal value of the pedestal evaluation value can be estimated more accurately in consideration of the influence of. When the elapsed time from the start of rotation is included in the feature amount group, it is preferable that the rotation direction of the door body 2 is also included in the feature amount group.

In the above processing example, it is possible to appropriately detect an abnormality in the configuration related to the rotation of the door body 2 by determining the presence or absence of an abnormality when the door body 2 is rotating. , The presence or absence of abnormality may be determined even when the door body 2 is stopped. In this case, for example, it is preferable that the feature amount group includes a value indicating whether or not the door body 2 is rotating.

In the abnormality detection device 40, as shown in FIG. 5, an inclinometer 12a may be used. The tilt meter 12a is attached to the pedestal 24 and acquires the tilt angle of the pedestal 24. As the inclinometer 12a, various inclinometers such as those using the inclination of the liquid level of the liquid held inside and those having an acceleration sensor can be used. The inclination angle of the pedestal 24 is input to the normal value estimation unit 41. In the normal value estimation unit 41, the inclination angle of the pedestal 24 is treated as the opening degree of the door body 2. Both the tilt angle of the pedestal 24 and the opening degree of the door body 2 may be used. With the inclinometer 12a, the opening degree of the door body 2 (inclination angle of the pedestal 24) can be acquired with a higher resolution than that of the inclinometer 12 in FIG.

In the radial gate 10, the tension of the wire rope 36 can be divided into a component perpendicular to the pedestal 24 and a component parallel to the pedestal 24, and the ratio of these components varies depending on the inclination angle of the pedestal 24. .. In the abnormality detecting device 40, by using the inclination meter 12a, the tension of the wire rope 36 can be accurately divided into a component perpendicular to the pedestal 24 and a component parallel to the pedestal 24. As a result, the normal value of the pedestal evaluation value can be estimated more accurately.

In the abnormality detection device 40, the load applied to the drive unit 31 of the switchgear 3 may be measured as a support load value. In one example, the load value acquisition unit 46 connected to the drive unit 31 measures the current value flowing through the motor of the drive unit 31 as the support load value. Such a support load value (current value) is also a value that substantially indicates the load applied to the wire rope 36. Further, the drive source in the drive unit 31 may be a hydraulic cylinder that performs linear motion, a hydraulic motor that performs rotary motion, or the like. In this case, in the load value acquisition unit 46, for example, the pressure in the hydraulic cylinder, the pressure of the hydraulic oil sent to the hydraulic cylinder or the hydraulic motor, or the current value of the electric motor used for pumping the hydraulic oil is the supporting load. Measured as a value. As described above, the support load value may be various kinds of values substantially indicating the load applied to the wire rope 36. When the drive source of the drive unit 31 is a hydraulic cylinder, the cylinder rod connected to the door body 2 serves as a connecting medium.

FIG. 6 is a block diagram showing a functional configuration of the abnormality detection device 40a according to the second embodiment. In the abnormality detection device 40a of FIG. 6, the evaluation value acquisition unit 47 in the abnormality detection device 40 of FIG. 3 is omitted, and the load value acquisition unit 46 is connected to the determination unit 42a. Further, the normal value estimation unit 41a estimates the normal value of the support load value, and the determination unit 42a compares the measured value of the support load value with the normal value. The other configurations of the abnormality detection device 40a are the same as those in FIG. 3, and the same configurations are designated by the same reference numerals.

FIG. 7 is a diagram showing a flow of processing in which the abnormality detection device 40a detects an abnormality in the radial gate 10. In the detection of the abnormality of the radial gate 10 in the abnormality detection device 40a, for example, the measured value of the support load value is acquired by the load value acquisition unit 46 when the door body 2 is rotated (step S21). The measured value of the support load value is input to the determination unit 42a. In the normal value estimation unit 41a, the side rubber submersion length is acquired based on the water level and the opening degree of the door body 2 in the same manner as in the above process (step S22).

In the normal value estimation unit 41a, the water level, the opening degree of the door body 2 and the side rubber submersion length are treated as feature quantity groups, and the feature quantity group is input to the estimator 411a which is a trained model. As a result, the normal value of the supporting load value is output from the estimator 411a (step S23). The estimator 411a in the abnormality detection device 40a is generated in advance by learning using a plurality of teacher data at a plurality of time points, as in the above processing example. Each teacher data here shows the measured value of the support load value acquired by the load value acquisition unit 46 at one time point, the water level, the opening degree of the door body 2 and the side rubber submersion length at that time point. It is a combination with the feature quantity group including.

As described above, in the door body 2, the moment due to the tension of the wire rope 36, the moment due to the weight of the door body 2, the moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911, and the bearing portion 23. The moment due to the frictional force between the rotating body and the bearing acts mainly. Further, the rotation of the door body 2 is low, and it can be considered that these moments are approximately balanced. The moment due to the weight of the door body 2 is affected by the opening degree of the door body 2. The moment due to the frictional force between the side watertight rubber 212 and the opening side surface 911 is affected by the submersion length of the side rubber. The moment due to the frictional force between the rotating body of the bearing portion 23 and the bearing is affected by the water level (water pressure). Therefore, in the estimator 411a generated by learning, the normal value of the supporting load value can be estimated from the feature amount group including the water level, the opening degree of the door body 2, and the submerged length of the side rubber.

In the determination unit 42a, the measured value of the support load value acquired by the load value acquisition unit 46 and the normal value of the support load value estimated by the normal value estimation unit 41a are compared, and the presence or absence of an abnormality in the radial gate 10 is found. It is determined (step S24). In the abnormality detection device 40a, the processes of steps S21 to S24 are repeated (step S25).

As described above, the load value acquisition unit 46 of the abnormality detection device 40a acquires the measured value of the support load value indicating the load applied to the wire rope 36. Further, the normal value estimation unit 41a estimates the normal value of the support load value based on the water level, the opening degree of the door body 2, and the submerged length of the side rubber. Then, the determination unit 42a compares the measured value of the supporting load value with the normal value, and determines whether or not there is an abnormality in the radial gate 10. Thereby, the abnormality in the radial gate 10 can be appropriately detected.

As described above, the side rubber submersion length is obtained from the water level and the opening degree of the door body 2. Therefore, even when the side rubber submersion length is omitted from the above feature quantity group, the water level and the water level and the opening of the door body 2 are used. Based on the opening degree of the door body 2, it is possible to estimate the normal value of the supporting load value with a certain degree of accuracy. Similarly, the water level and the opening degree of the door body 2 may be omitted from the feature quantity group, and in this case as well, the normal value of the supporting load value is obtained with a certain degree of accuracy by using the side rubber submersion length. It is possible to estimate. Even when the feature amount group includes only the side rubber submersion length, it can be said that the normal value of the supporting load value is substantially estimated based on the water level and the opening degree of the door body 2.

Similar to the abnormality detection device 40 of FIG. 3, in the abnormality detection device 40a of FIG. 6, it is preferable that the rotation direction of the door body 2 is included in the feature amount group. Further, the discharge flow rate, the (current) rotation direction of the door body 2, the previous rotation direction, the elapsed time from the start of rotation of the door body 2, and the like may be included in the feature quantity group. The evaluation value acquisition unit 47 described above may be provided, and the pedestal evaluation value may be included in the feature quantity group. Further, the presence or absence of an abnormality may be determined even when the door body 2 is stopped. The inclinometer 12a of FIG. 5 may be used, and the load applied to the drive unit 31 of the switchgear 3 may be measured as a support load value. The normal value of the supporting load value may be estimated without using the estimator 411a.

The abnormality detection devices 40, 40a, the abnormality detection method, and the floodgate system 1 can be modified in various ways.

Depending on the design of the radial gate 10, side watertight rubber may be provided on each opening side surface 911 at a position where the side portion 211 of the skin plate 21 comes into contact with the rotation of the door body 2. That is, the side watertight rubber may be attached to either the side portion 211 or the opening side surface 911.

As shown in FIG. 2, when a part (upper part) of the side watertight rubber 212 does not come into contact with the opening side surface 911 due to the rotation of the door body 2 in which the skin plate 21 rises, the opening side surface 911 in the side watertight rubber 212 The length of the portion that does not come into contact with the feature amount group may be included in the feature amount group.

The radial gate 10 may be a tensile radial gate in which water flows from the support portion 23 side to the skin plate 21 side.

The above-described embodiments and configurations in the respective modifications may be appropriately combined as long as they do not conflict with each other.

1 Water gate system 2 Door body 10 Radial gate 12a Incliner 21 Skin plate 23 Bearing part 24 Pedestal 31 Drive part 36 Wire rope 40,40a Abnormality detection device 41,41a Normal value estimation part 42,42a Judgment part 46 47 Evaluation value acquisition part 211 (of skin plate) Side part 212 Side part Watertight rubber 411,411a Estimator 911 Opening side surface S11 to S15, S21 to S25 Step

Claims (10)

An anomaly detection device that detects anomalies in radial gates.
In the radial gate, the door body and the drive unit are connected by a connecting medium, and a load value acquisition unit that acquires a value indicating the load applied to the connecting medium as a support load value, and a load value acquisition unit.
In the door body, the skin plate and the bearing portion are connected by a pedestal, and an evaluation value acquisition unit that acquires a value indicating stress applied to the pedestal as a pedestal evaluation value, and an evaluation value acquisition unit.
A normal value estimation unit that estimates the normal value of the pedestal evaluation value based on the water level at the installation position of the radial gate, the opening degree of the door body, and the support load value.
A determination unit for determining the presence or absence of an abnormality in the radial gate by comparing the pedestal evaluation value acquired by the evaluation value acquisition unit with the normal value estimated by the normal value estimation unit.
Anomaly detection device characterized by being equipped with. The abnormality detection device according to claim 1.
Each side of the skin plate comes into contact with the door stop via the watertight rubber,
An abnormality detection device, wherein the normal value estimation unit estimates the normal value when the door body rotates based on the length of the watertight rubber that receives water pressure. The abnormality detection device according to claim 1 or 2.
An abnormality detection device, wherein the normal value estimation unit estimates the normal value when the door body rotates based on the flow rate of water passing through the radial gate. The abnormality detection device according to any one of claims 1 to 3.
An abnormality detection device, wherein the normal value estimation unit estimates the normal value based on the elapsed time from the start of rotation of the door body when the door body is rotated. The abnormality detection device according to any one of claims 1 to 4.
An anomaly detection device further comprising an inclinometer attached to the pedestal. The abnormality detection device according to any one of claims 1 to 5.
The pedestal evaluation value acquired by the evaluation value acquisition unit at one time point in the normal state, and the feature amount group substantially including the water level, the opening degree, and the support load value at the one time point. It is characterized in that the normal value estimation unit estimates the normal value of the pedestal evaluation value using a trained model generated by learning using a plurality of teacher data at a plurality of time points using the combination as teacher data. Anomaly detection device. An anomaly detection device that detects anomalies in radial gates.
In the radial gate, the door body and the drive unit are connected by a connecting medium, and a load value acquisition unit that acquires a value indicating the load applied to the connecting medium as a support load value, and a load value acquisition unit.
A normal value estimation unit that estimates the normal value of the support load value based on the water level at the installation position of the radial gate and the opening degree of the door body.
A determination unit for determining the presence or absence of an abnormality in the radial gate by comparing the support load value acquired by the load value acquisition unit with the normal value estimated by the normal value estimation unit.
Anomaly detection device characterized by being equipped with. It ’s a floodgate system.
Radial gate and
The abnormality detection device according to any one of claims 1 to 7, which detects an abnormality in the radial gate.
A floodgate system characterized by being equipped with. Anomaly detection method for detecting anomalies in radial gates
a) In the radial gate, the door body and the drive unit are connected by a connecting medium, and a step of acquiring a value indicating a load applied to the connecting medium as a supporting load value, and
b) A step in which the skin plate and the bearing portion are connected by a pedestal in the door body, and a value indicating the stress applied to the pedestal is acquired as a pedestal evaluation value.
c) A step of estimating a normal value of the pedestal evaluation value based on the water level at the installation position of the radial gate, the opening degree of the door body, and the supporting load value.
d) A step of determining the presence or absence of an abnormality in the radial gate by comparing the pedestal evaluation value acquired in the step b) with the normal value estimated in the step c).
Anomaly detection method comprising. Anomaly detection method for detecting anomalies in radial gates
a) In the radial gate, the door body and the drive unit are connected by a connecting medium, and a step of acquiring a value indicating a load applied to the connecting medium as a supporting load value, and
b) A step of estimating a normal value of the supporting load value based on the water level at the installation position of the radial gate and the opening degree of the door body.
c) A step of determining the presence or absence of an abnormality in the radial gate by comparing the supporting load value acquired in the step a) with the normal value estimated in the step b).
Anomaly detection method comprising.

Images