Valve state diagnostic device and state diagnostic method

Abstract

A valve state diagnostic device detailedly indicates a change in the operating state of a valve and a change prediction. In a valve state diagnostic device, when an angle sensor detects an operation of a valve, a processor executes the following steps: a step of acquiring an angle profile related to the operation of the valve from the angle sensor; a step for performing histogram processing on the acquired angle profile, averaging a plurality of previous and subsequent operations, and performing relative frequency processing; a step for calculating the similarity between the histogram of the angular profile after the relative frequency conversion and the histogram of the angular profile at the initial stage of operation by means of a histogram crossing method; plotting the calculated similarity of the angle profiles in a two-dimensional plane in which the similarity of the angle profiles of the opening operation is taken on the horizontal axis and the similarity of the angle profiles of the closing operation is taken on the vertical axis; and a step of performing clustering processing on plot data of the similarity of the angle contour generated by the step of repeating plotting a plurality of times.

Description

Technical Field

[0001] The present invention relates to a valve state diagnosis device and a state diagnosis method. Background Art

[0002] Valves are used in various parts of various workshops and factories. After the valves are assembled into workshop or factory equipment, they are required to operate continuously and stably for a long time. However, most of the assembled valves are arranged in positions, parts, and environments where operators and inspectors are difficult to directly access. In addition, it is difficult for operators and inspectors to grasp the current state of the valve only through external visual inspection, etc., and it is also difficult to accurately predict the future state change of the valve.

[0003] It is required to appropriately diagnose valves that are continuously used for a long time as part of various equipment, so as to accurately grasp the current state of the valves, and accurately predict the state change of the valves. Further, through these grasps and predictions, an accurate equipment maintenance plan can be formulated and potential failures can be prevented.

[0004] Patent Documents 1, 2, and 3 disclose valve diagnosis devices and methods. First, in the "Automatic Valve Diagnosis System and Automatic Valve Diagnosis Method" disclosed in Patent Document 1, data related to the valve detected by various sensors is compared with model data for failure determination. In this method, multiple sensors are required for diagnosing the valve state, and the system may be enlarged. In addition, regarding valve failures, it is difficult to determine failures when there is no model data obtained with the same sensor structure. Further, in failure prediction, it only presents information such as "it may leak in the near future", and cannot present specific information, for example, a specific time reference until failure occurs.

[0005] In the "Regulator, Regulation Method, and Regulation System" and the "Actuator Operation Condition Detection Device and Operation Condition Detection Method" disclosed in Patent Documents 2 and 3, failure prediction is performed by comparing the opening and closing time of the valve with past data. This method only observes the change in the opening and closing time, and does not observe the change in the subtle operation condition of the valve. Therefore, it is difficult to determine the failure location during the occurrence and prediction of failures. When a valve fails, it mostly goes through a process of gradual performance degradation. Therefore, even if one wants to detect valve abnormalities as early as possible, if the abnormality is not detected until it has expanded to cause a change in the opening and closing time, it is a bit too late. Prior Art Documents Patent Documents

[0006] Patent Document 1: Japanese Patent Application Laid-Open No. 2018-73154 Patent Document 2: Japanese Patent Application Laid-Open No. 2019-191760 Patent Document 3: Japanese Patent Application Laid-Open No. 2004-169887 Summary of the Invention Technical Problem to be Solved by the Invention

[0007] It is required that the valve state diagnosis device diagnose the state of the valve with a simple structure and show in detail the change in the operation state of the valve. In addition, it is required that the valve state diagnosis device show not only abstract prediction information but also information urging the operator's specific actions, such as the specific remaining time predicted until a failure occurs and the remaining number of opening and closing operations. Further, it is thus required that the valve state diagnosis device be able to formulate an accurate equipment maintenance plan and prevent future failures before they occur. Technical Solution for Solving the Technical Problem

[0008] The valve state diagnosis device of the present invention includes an angle sensor that detects the operating angle of the valve and a processor. In this valve state diagnosis device, when the angle sensor detects the operation of the valve, the processor performs the following steps: (1) A step of obtaining an angle profile related to the operation of the valve from the angle sensor; (2) A step of histogramming the obtained angle profile, averaging the operations before and after, and performing relative frequency conversion; (3) A step of calculating the similarity between the histogram of the relatively frequency-converted angle profile and the histogram of the angle profile at the initial stage of operation by the histogram intersection method; (4) A step of plotting the calculated similarity of the angle profile in a two-dimensional plane where the similarity of the angle profile in the opening operation is taken on the horizontal axis and the similarity of the angle profile in the closing operation is taken on the vertical axis; and (5) A step of performing clustering processing on the plotted data of the similarity of the angle profile generated by repeatedly performing the plotting step of the above (4) multiple times. Advantages of the Invention

[0009] The valve state diagnosis device of the present invention can diagnose the state of the valve with a simple structure and show in detail the change in the operation state of the valve and the change prediction to the operator. Further, the valve state diagnosis device of the present invention can thus achieve the formulation of an accurate equipment maintenance plan and prevent failures before they occur. Description of the Drawings Figure 1 It is a block diagram showing the structure of the valve state diagnosis device according to the embodiment. Figure 2 It is a perspective view and a partial enlarged view of the valve state diagnosis device and the cylinder block according to the embodiment. Figure 3 It is an overall structure diagram of the valve state diagnosis device and an external network according to the embodiment. Figure 4 It is a flowchart of the process for valve state diagnosis in the valve state diagnosis device according to the embodiment. Figure 5 It is an example of an angular profile related to the operation of the valve. Figure 6 It is a diagram showing the expansion of the angular profile into a frequency distribution table and the creation of a histogram. Figure 7 It is a diagram showing the relative frequency conversion of the histogram of the angular profile. Figure 8 It is a diagram showing the histogram intersection method for calculating the similarity between angular profiles. Figure 9 It is a diagram showing an example of a plot of the similarity of angular profiles. Figure 10 It is a diagram showing the plot data of the similarity of the angular profiles subjected to clustering processing, which is used to grasp the current state of the valve. Figure 11 It is a diagram showing the clustering process for the plot of the similarity of angular profiles to confirm the change in the product life cycle of the valve. Figure 12 It is a coordinate diagram of the plot of the similarity of the angular profiles subjected to clustering processing for predicting the state of the valve based on prior data, and a table of the existence ratios of each period of the product life cycle obtained in advance through experiments. Figure 13 It is a diagram showing the use of the plot of the similarity of angular profiles to evaluate and predict the state of the valve by the number of clusters. Figure 14 It is a diagram showing an example of the calculation of the number of clusters. Figure 15 It is a diagram showing a prediction example of the transition timing of the product life cycle of the valve. Figure 16 It is an example of the screen of a smartphone that notifies the results of the valve state prediction and diagnosis. Detailed Embodiment

[0011] Hereinafter, the embodiment will be described in detail with appropriate reference to the accompanying drawings. However, sometimes a too detailed description may be omitted. For example, sometimes the detailed description of well-known matters or the repeated description of substantially the same structure may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art.

[0012] It should be noted that the inventors of the present invention provide the accompanying drawings and the following description in order to enable those skilled in the art to fully understand the present invention, and do not intend to limit the subject matter described in the scope of the patent claim by these.

[0013] 1. Origin of the present invention Valves are used in various parts of various workshops and factories. After the valves are assembled into the equipment in the workshops and factories, they are required to operate continuously and stably for a long time.

[0014] However, most of the assembled valves are arranged in positions, parts, and environments where operators and inspectors are difficult to directly access. In addition, it is difficult for operators and inspectors to grasp the current state of the valves only by visual inspection from the outside of the equipment, and it is difficult to accurately predict the future state changes of the valves.

[0015] In addition, there is a concern that due to the aging of the workshops, the aging or retirement of senior maintenance personnel and managers, there is insufficient inheritance of management skills, etc., and the risk of major accidents will increase in the future.

[0016] On the other hand, since multiple valves can be arranged in a complex environment, it is required to grasp the current state related to the valves and predict the future state changes of the valves with a simple structure as much as possible.

[0017] In addition, it is required to construct a device for valve inspection and diagnosis that can not only show abstract prediction information but also show information urging the operator's specific actions, such as the specific remaining time until a failure occurs and the remaining number of opening and closing times.

[0018] Moreover, it is also required to construct a device that meets the need to always monitor the state of unmanned equipment, a device that can confirm the valve state remotely, and a device that can monitor multiple valves together.

[0019] The present invention provides a valve state diagnosis device and a state diagnosis method that can diagnose the valve state with a simple structure and can represent in detail the changes in the valve operation status and the change prediction to the operator.

[0020] 2. (Embodiment) Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

[0021] 2.1. Structure of the valve state diagnosis device Figure 1 It is a block diagram showing the structure of the valve state diagnosis device 2 and the surrounding devices according to the embodiment. As will be described later Figure 2 As will be described later, the valve state diagnosis device 2 is provided on the upper part of the cylinder 12 that is especially used to move the valve core (valve) in the valve part 22.

[0022] The valve state diagnosis device 2 includes an angle sensor 4, a processor 6, a memory 8, and a communication unit 10.

[0023] The angle sensor 4 is a sensor that detects the rotational movement of the magnet 14 provided at the upper part of the cylinder 12 in a non-contact manner, and is mainly composed of TMR (Tunnel Magnetoresistance Element) / GMR (Giant Magnetoresistance Element) elements. It should be noted that the magnet 14 is configured to rotate corresponding to the movement of the cylinder 12 and further to the rotational movement of the valve in the valve unit 22. Therefore, the angle sensor 4 detects the operating angle of the valve. The magnet 14 is assumed to be in a ring shape, rod shape, fan shape, etc., but it can also be other shapes. In addition, the angle sensor 4 here is a magnetic angle sensor paired with the magnet 14, but the angle sensor 4 is not limited to a magnetic angle sensor. For example, it can also be an angle sensor based on a potentiometer, an angle sensor based on a rotary encoder, a gyro sensor as an angular velocity sensor, etc.

[0024] The processor 6 is composed of a CPU (Central Processing Unit). Each function of the valve state diagnosis device 2 according to the present embodiment is realized by the processor 6 executing various programs. It should be noted that each of these functions can be realized by an ASIC (Application Specific Integrated Circuit) or the like, or can also be realized by a combination of an ASIC and a CPU loaded with various programs.

[0025] The memory 8 is a storage unit capable of rewriting data inside the valve state diagnosis device 2, and is composed of, for example, a RAM (Random Access Memory) including a plurality of semiconductor storage elements. The memory 8 temporarily stores specific computer programs, variable values, parameter values, etc. when the processor 6 executes various processes. The memory 8 stores, for example, (valve) opening / closing trend data, angle profile data, histogram data, operation data, and various setting data as intermediate data or result data of valve state diagnosis.

[0026] Here, the (valve) opening / closing trend data refers to a set of information indicating the operating condition of the valve, such as the action date and time of the valve / the time required for the action / the action direction / the initial angle / the final angle, etc. The angle profile data refers to two-dimensional data obtained by plotting with the vertical axis being the angle of the valve (from fully open to fully closed or from fully closed to fully open) and the horizontal axis being the time for one action of the valve, and is data indicating how the valve operates during the opening / closing action.

[0027] It should be noted that the memory 8 may include a so-called ROM (Read Only Memory). A computer program for implementing the processing of the valve state diagnosis device 2 described below is pre-stored in the ROM. The processor 6 reads the computer program from the ROM and expands it in the RAM, and the processor 6 can execute the computer program.

[0028] The communication unit 10 is a circuit for performing wired and wireless communications with the outside, and is an interface circuit including various network terminals, USB terminals, etc., capable of outputting data to the outside and acquiring data from the outside. Interim data and result data related to valve state diagnosis are sent to, for example, a PC (Personal Computer) 16 via the wired communication function of the communication unit 10. On the other hand, similarly, the interim data and result data related to valve state diagnosis are sent to, for example, a smart phone 18 via the wireless communication function of the communication unit 10. It should be noted that the smart phone 18 may also be a tablet terminal or the like.

[0029] It should be noted that the interim data and result data related to valve state diagnosis sent to the PC (Personal Computer) 16 and the smart phone 18 can be sent to a cloud server 20 via an external network such as the Internet 21.

[0030] In the present embodiment, it is assumed that each process related to valve state diagnosis is performed by the processor 6 of the valve state diagnosis device 2. The PC 16 and the smart phone 18 only represent the results of each process related to valve state diagnosis, that is, the current state and the prediction of the state related to the valve.

[0031] It should be noted that it may also be in the following manner: the data collected by the angle sensor 4 is sent approximately directly to the PC 16 and the smart phone 18, and each process related to valve state diagnosis is performed by these PC 16 and smart phone 18. In this case, the main operation of the valve state diagnosis device 2 is to collect the angular profile data related to the operating angle of the valve, temporarily store it, and then send it to the PC 16 and the smart phone 18.

[0032] In addition, it may also be in the following manner: the data collected by the angle sensor 4 is sent approximately directly to the PC 16 and the smart phone 18, and further, these data are sent from the PC 16 and the smart phone 18 to the cloud server 20 via the Internet 21, and each process related to valve state diagnosis is performed in the cloud server 20. In this case, the main operation of the valve state diagnosis device 2 is also to collect the angular profile data related to the operating angle of the valve, temporarily store it, and then send it to the PC 16 and the smart phone 18.

[0033] Figure 2 (1) of which is a perspective view of the valve state diagnosis device 2 and the cylinder 12 according to the embodiment, and Figure 2(2) is a partial enlarged view of the valve state diagnosis device 2 and the cylinder 12 involved in the embodiment. In Figure 2 In the partial enlarged view of (2), the relationship between the valve state diagnosis device 2 and the annular magnet 14 provided at the upper part of the cylinder 12 is particularly shown. Further, as Figure 2 shown in the perspective view of (1), the valve state diagnosis device 2 is configured to include a USB connector 24, a power supply, and an RS485 connector 26, which are respectively connected to a USB communication cable 28 and an RS485 communication cable 30.

[0034] Figure 3 is an overall structure diagram of the valve state diagnosis device 2 and the external network involved in the embodiment. As Figure 3 shown, the valve state diagnosis device 2 provided at the upper parts of the valve portion 22 and the cylinder 12 is connected to the PC 16 via a USB communication cable 28. In addition, the valve state diagnosis device 2 is connected to the PC 16 via an RS485 communication cable 30, a USB-RS485 converter 32, and a USB communication cable 28. Further, the valve state diagnosis device 2 is connected to the smart phone 18 via Bluetooth (registered trademark) 34. The PC 16 and the smart phone 18 are connected to the cloud server 20 via the Internet 21.

[0035] 2.2. (Operation of Valve State Diagnosis Device) Figure 4 is a flowchart of the process for valve state diagnosis in the valve state diagnosis device 2 involved in the embodiment. After the power of the valve state diagnosis device 2 is turned on (step S04), an initialization process (step S06) is performed as an initial setting of the device. Then, as long as an opening action or a closing action of the valve occurs, the valve state diagnosis device 2 repeatedly executes the main program composed of the current situation grasping process (program S100) and the state prediction process (program S200).

[0036] It should be noted that the opening action of the valve refers to the action of the valve (valve core) from the closed state (i.e., fully closed) to the open state (i.e., fully open), and the closing action of the valve refers to the action of the valve (valve core) from the open state (i.e., fully open) to the closed state (i.e., fully closed).

[0037] 2.2.1. Current Situation Grasping Process In the current situation grasping process (program S100), by the occurrence of an opening action or a closing action of the valve, first, the valve state diagnosis device 2 acquires opening and closing trend data (step S08).

[0038] Next, the valve state diagnosis device 2 acquires an angular profile related to the action of the valve (step S10). Figure 5This is an example of the angular profile related to the operation of the valve. It should be noted that the valve state diagnosis device 2 can also be configured to execute step S08 and step S10 in parallel, while acquiring the opening / closing trend data and the angular profile data.

[0039] In Figure 5 the valve showing the angular profile during operation is a center-type rubber-sealed valve. Hereinafter, similarly for the example of the valve adopted, unless otherwise specified, it is a valve related to a center-type rubber-sealed valve. The application of the valve state diagnosis device 2 according to the present embodiment is of course not limited to the center-type rubber-sealed valve. The valve state diagnosis device 2 of the present embodiment can also be applied to an eccentric-type valve, for example.

[0040] It should be noted that the valve state diagnosis device 2 of the present embodiment can be applied to a valve for acquiring the angular profile near 90 degrees, such as a butterfly valve and a ball valve. In addition, it can also be similarly applied to a multi-turn opening / closing valve capable of acquiring the angular profile, such as a gate valve and a globe valve.

[0041] Next, the valve state diagnosis device 2 analyzes the angular profile (step S12). In this step of analyzing the angular profile, (a) The angular profile is expanded in a frequency distribution table and made into a histogram. (b) The histogram is averaged. (c) Relative frequency conversion is performed.

[0042] Figure 6 is a diagram showing the expansion of the angular profile into a frequency distribution table and the creation of a histogram. (a) In the expansion into a frequency distribution table and the creation of a histogram, first, the angular profile data is divided according to each given angle (class) ( Figure 6 (1) of Figure 6 ), and a frequency distribution table is created by summing based on the class ( Figure 6 (2) of

[0043] (b) In the averaging of the histogram, the angular profiles of multiple times before and after the operation of the object are totaled to create a frequency distribution, resulting in a histogram that equalizes the changes in the periodic or sudden operating states.

[0044] (c) In the relative frequency conversion, in all classes (angle ranges), the ratio of the frequency of each class to the whole is calculated to create a histogram. Figure 7It is a graph showing the relative frequency of the histogram representing the angular profile. Figure 7 (1a) of Figure 7 is a frequency distribution table of the number of data for one operation (opening operation) of the valve. Figure 7 (1b) of Figure 7 is Figure 7 a graph obtained by histogramming the frequency distribution histogram shown in (1a) of Figure 7 . Figure 7 (2a) of Figure 7 is Figure 7 a frequency distribution table obtained by relative frequencying the frequency distribution shown in (1a) of Figure 7 . Figure 7 (2b) of Figure 7 is Figure 7 a graph obtained by histogramming the frequency distribution shown in (2a) of Figure 7 . Figure 7 The frequency distributions and histograms illustrated in (2a) and (2b) of Figure 7 represent the proportion of time that the valve core exists in each angular range (grade) during one operation of the valve. (c) Relative frequencying is performed to compare the frequency distribution and histogram related to a certain operation in the valve with those related to other operations in the same valve. Considering that the time required for each operation of the valve changes.

[0045] Next, the valve state diagnosis device 2 diagnoses the angular profile (step S14). In this step of diagnosing the angular profile, the similarity between the histogram of the angular profile and the histogram of the angular profile at the initial stage of operation is calculated. In the calculation of the similarity, various methods can be adopted. For example, the histogram intersection method can be used. Figure 8 is a graph showing the histogram intersection method for calculating the similarity between angular profiles. The similarity of the histogram intersection method is calculated by comparing the two relative frequencyed histograms for each grade and summing the smaller frequency. In Figure 8 , since the size of the overlapping part between histogram A and histogram B is 0.45, "similarity: 0.45" is calculated. Therefore, when the shapes of the histograms being compared are very different, the similarity becomes low. That is, if the similarity is small, it means that the shape of the angular profile has changed significantly.

[0046] Next, the valve state diagnosis device 2 performs a similarity analysis of the histogram of the angular profile (step S16). In this step of analyzing the similarity of the histogram of the angular profile, in a two-dimensional plane where the similarity of the angular profile during the opening operation is taken on the horizontal axis (x-axis) and the similarity of the angular profile during the closing operation is taken on the vertical axis (y-axis), the similarity of the angular profile for each given number of opening and closing operations is plotted. The given number of opening and closing operations is, for example, 1000 times for the opening operation and 1000 times for the closing operation. Figure 9 (1), (2), and (3) of Figure 9 are graphs showing three examples of the plots of the similarity of the angular profile. Figure 9 (1), (2), and (3) of Figure 9 are plots related to three different valves. All are plotted for every 1000 opening and closing operations.

[0047] In the coordinate graph of the plot of the similarity of the angular profile, "plotted every 1000 opening / closing operations" is the same in Figures 10 to 14 as well.

[0048] It should be noted that in the step of analyzing the similarity of the histogram of the angular profile (step S16), the similarity of the angular profile is plotted at given intervals of the opening / closing operations. Therefore, for the opening / closing operations of the valve other than the given intervals of the opening / closing operations, this step S16 may not be performed.

[0049] Next, the valve state diagnosis device 2 confirms the current state in the product life cycle of the valve (step S18). In this step of confirming the current state in the product life cycle, clustering processing is performed on the plot data of the similarity of the angular profile generated by repeatedly performing step S16 multiple times, and it is speculated which period of the product life cycle the latest plot data corresponds to. Figure 10 It is a graph showing the plot data of the similarity of the angular profile that has undergone clustering processing and is used to grasp the current state of the valve. Figure 10 Among the plot data of the similarity of the angular profile that has undergone clustering processing shown, as the product life cycle of the center type rubber seal valve, five periods, namely, "new product period", "running-in period", "stable period", "deterioration period", and "failure period" are assumed. Figure 10 It shows that the current state (i.e., "latest data") is in the deterioration period.

[0050] That is, in the whole of the processing for valve state diagnosis in the valve state diagnosis device 2 according to the present embodiment, based on the experimental data related to the opening / closing operations repeatedly performed on the valve, the product life cycle is assumed for each product type of the valve. As described above, the product life cycle of the center type rubber seal valve is assumed to be five periods (as follows). In particular, the number of these periods is used as the maximum number of clusters assumed in the clustering processing. (1) New product period It is a period in which each component is in a new product state and before the components are run in with each other. It is a period in which substances such as grease temporarily attached during assembly exist. (2) Running-in period It is a period in which each component is in a running-in state with each other. During this period, substances such as grease temporarily attached during assembly flow out. (3) Stable period It is a period in which each component can operate normally. It is a period in which wear occurs due to the contact and sliding of the components but is in a state of stable wear. During this period, a state in which sealing can be performed at the maximum allowable pressure is maintained. (4) Deterioration period It is the period when the deterioration rate of each component accelerates. For example, it is the period when starting from damage, burrs, etc. generated on the component surface, it gradually changes to a state where the target component is severely worn. In this stage, sealing cannot be performed at the allowable maximum pressure, and the pressure that can be sealed continuously decreases. (5) Failure period It is the period when one of the components constituting the valve cannot function, for example, it is the period when it is in a state where one or more of the following symptoms occur. · The valve stem breaks. · The clearance between the valve stem and the valve core becomes larger, and the valve core does not rotate to the fully closed position. · The sealing ring wears, and it cannot be sealed even at about 0.1 MPa (megapascal).

[0051] It should be noted that the details of the "clustering process" will be described in the subsequent "state prediction process".

[0052] Furthermore, in the step of confirming the current situation in the product life cycle of the valve (step S18), it is also possible to notify the outside of the faults with high possibility among the faults assumed in the corresponding product life cycle based on the opening / closing trend data and the angle profile data. The assumed faults and the levels of possibility here can be determined, for example, based on the experimental data related to the opening / closing operations repeatedly performed on various center-type rubber-sealed valves.

[0053] It should be noted that step S16 is implemented at given intervals of the number of opening / closing operations, so step S18 can also be performed at corresponding given intervals of the number of opening / closing operations.

[0054] 2.2.2. State prediction process In the state prediction and grasping process (program S200), the valve state diagnosis device 2 performs the following respective processes based on the information obtained from the analysis of the similarity with the angle profile. However, the following respective processes can also be performed at given intervals of the number of opening / closing operations corresponding to the plotting of the similarity with the angle profile, similar to steps S16 and S18. · Confirmation of product life cycle change (step S20). · State prediction based on prior data (step S22). · State prediction based on the evaluation of the number of clusters (step S24). · Notification of state prediction / diagnosis results (step S26).

[0055] In the state prediction process (program S200), first, the valve state diagnosis device 2 confirms the change in the product life cycle (step S20). In this step of the confirmation process of the product life cycle change, a coordinate graph for plotting the similarity of the angle profile (for example, refer toFigure 9 ) clustering process. The clustering process in the valve state diagnosis device 2 involved in the embodiment is not particularly limited. For example, PAM (Partition Around Medoids) is used. That is, the valve state diagnosis device 2 assumes the number of periods in the product life cycle as the maximum number of clusters, and determines the current number of clusters in the coordinate graph of the plotted similarity of the angular profiles based on the calculation of the gap statistic using PAM. On this basis, the valve state diagnosis device 2 confirms whether the number of clusters has changed, for example, whether it has increased.

[0056] In the above calculation of the gap statistic using PAM, the evaluation value obtained based on the evaluation function is obtained. Here, the evaluation function is defined as the gap (absolute value of the difference) between the "average distance within the cluster in A" and the "average distance within the cluster in B" based on the specified number of clusters when A is the data to be processed and B is the uniformly distributed data. The value of this gap is the evaluation value obtained based on the evaluation function. For example, assuming the number of clusters is from 1 to 5, the absolute value of the difference between the average distance within the cluster (from 1 to 5) in the data to be processed and the average distance within the cluster (from 1 to 5) in the uniformly distributed data is calculated, and the number of clusters with the highest absolute value of this difference (i.e., the evaluation value obtained based on the evaluation function) is determined as the current number of clusters (see Figure 14 ).

[0057] Figure 11 It is a diagram showing the clustering process of the coordinate graph of the plotted similarity of the angular profiles to confirm the change in the product life cycle of the valve. In Figure 11 the shown clustering process, it is confirmed that the number of clusters changes from 2 to 3. This indicates that it is confirmed that the product life cycle of the valve is entering the "stable period" or has entered the "stable period".

[0058] It is required to predict the transition period to the subsequent periods (here, the "deterioration period" and the "failure period") of the product life cycle by confirming that the product life cycle is entering the "stable period". Therefore, the valve state diagnosis device 2 then performs "state prediction based on prior data (step S22)" and "state prediction based on cluster number evaluation (step S24)".

[0059] Next, the valve state diagnosis device 2 performs state prediction based on prior data (step S22). In this step of the state prediction process based on prior data, based on the existence ratio of each period of the product life cycle obtained in advance through experiments for each product type of the valve, the position of the latest plotted data in the current product life cycle of the valve for the latest plotted data is grasped.

[0060] Figure 12The table on the right shows an example of the existence ratio of each period of the product life cycle of the center type rubber seal valve obtained through experiments in advance. Grasp the position where the latest (current) plotted data exists within the period (i.e., stage) of the product life cycle corresponding to the current number of clusters calculated based on the clustering process in step S20. For example, Figure 12 The coordinate diagram on the left shows that the current number of clusters is 3, that is, the center type rubber seal valve of the object is in the "stable period". According to Figure 12 the number of plots in the coordinate diagram on the left and Figure 12 the existence ratio of each period in the table on the right, for example, it can be grasped that the latest (current) plotted data is near "43%" within "45%" of the "stable period". Then, the valve state diagnosis device 2 can determine that "the period when the number of clusters changes from 3 to 4 is approaching".

[0061] In step S22, based on the table of the existence ratio of each period of the product life cycle obtained through experiments in advance for each product type of the valve, grasp the position within the period of the product life cycle, and perform state prediction processing based on prior data. At this time, data related to the cumulative movement angle of the valve spool can also be used assistively. Here, the "cumulative movement angle of the valve spool" is calculated from the cumulative opening and closing trend data.

[0062] Next, the valve state diagnosis device 2 performs state prediction processing based on the evaluation of the number of clusters (step S24). In this state prediction processing based on the evaluation of the number of clusters, the gap statistic is calculated. That is, the gap statistic is calculated for the number of clusters assumed for the product type of the target valve. For example, if it is a center type rubber seal valve, the gap statistic is calculated for the number of clusters from 1 to 5. As shown in step S20, the number of clusters when the evaluation value obtained based on the evaluation function in the gap statistic calculation for each number of clusters is the highest is set as the final current number of clusters.

[0063] Furthermore, based on the evaluation value obtained based on the evaluation function in the gap statistic calculation for each number of clusters and the trend of the difference between the evaluation values, predict the trend of the period of the product life cycle of the target valve. Specifically, by calculation, for example, in the form of the cumulative opening and closing action times, using the predicted curve of the transition period, calculate the timing when the evaluation value of "the current number of clusters" and the evaluation value of "the number of clusters in the next stage (period)" are reversed, that is, the timing of transitioning to the next period of the product life cycle.

[0064] Figure 13 It is a coordinate diagram showing the plot using the similarity of the angle profile to predict the state of the valve through the evaluation of the number of clusters.

[0065] Furthermore, Figure 14It is a diagram showing an example of calculating the number of clusters. It shows the calculation of the number of clusters based on the gap statistic, where PAM (Partition around medoids) is used in the clustering process. Figure 14 On the upper left is a coordinate diagram of the plot of the similarity of the angular profiles every 1000 opening / closing operations at the moment when the opening / closing operation of a certain center-type rubber seal valve reaches 200000 times. Figure 14 On the upper right is a coordinate diagram of the evaluation values (y-axis) obtained based on the evaluation function in the gap statistic calculation for the number of clusters 1, 2, 3, 4, 5 (x-axis). The evaluation value is the highest when it is "4", and the final current number of clusters is determined to be "4".

[0066] Figure 14 On the lower left is (the same as the Figure 14 upper valve) a coordinate diagram of the plot of the similarity of the angular profiles every 1000 opening / closing operations at the moment when the opening / closing operation of the center-type rubber seal valve reaches 250000 times. Figure 14 On the lower right is a coordinate diagram of the evaluation values (y-axis) obtained based on the evaluation function in the gap statistic calculation for the number of clusters 1, 2, 3, 4, 5 (x-axis). The evaluation value is the highest when it is "5", and the final current number of clusters is determined to be "5".

[0067] Furthermore, Figure 15 It is a diagram showing a prediction example of the transfer timing of the product life cycle, that is, the number of clusters. Figure 15 On the left, it shows the evaluation values obtained based on the evaluation function in the gap statistic calculation for the number of clusters 2 and 3, and the difference in evaluation values, at the moments when the opening / closing times in the center-type rubber seal valve are 12500 times, 126000 times,..., 133000 times,.... In this table, predicted values are shown after the opening / closing times reach 134000 times. Figure 15 On the right is a predicted curve of the transition period generated based on the difference in evaluation values up to 133000 times. The predicted curve of the transition period here is represented by the following formula. (Equation 1) y = 7E-22× 5 -5E-16×4 + 1E-10x 3 -2E-05x 2 +1.174x - 31256

[0068] In Figure 15 the table on the left, it shows that the difference in evaluation values decreases as the opening / closing times increase. In Figure 15In the coordinate graph on the right, the x-axis value (number of opening / closing operations) at the moment when the predicted value (difference in evaluation values) for the y-axis becomes 0 is the timing of the product life cycle transition. That is, at the moment when the number of opening / closing operations is 133,000, it is predicted that the transition will occur from cluster number 2 (running-in period) to cluster number 3 (stable period) when the number of opening / closing operations is approximately 140,000.

[0069] Next, the valve state diagnosis device 2 performs a result notification process for state prediction and diagnosis via the communication unit 10 (step S26). That is, the valve state diagnosis device 2 sends the information generated by diagnosing the valve to an external computer via the communication unit 10. In this step of notifying the results of state prediction and diagnosis, the valve state diagnosis device 2 sends, for example, the following content to the PC 16 and the smartphone 18. The PC 16 and the smartphone 18 display the following content. (1) Regarding the current stage (period) of the product life cycle of the valve (2) The number of opening / closing operations until the next stage (period) (3) The predicted time to reach the next stage

[0070] Furthermore, the valve state diagnosis device 2 can also send the countermeasure policy that the operator should take when reaching the next stage to the PC 16 and the smartphone 18. This countermeasure policy can be preset according to the stage and the cumulative number of opening / closing operations, and the content of this countermeasure policy is pre-stored in the memory 8.

[0071] Figure 16 is an example of the screen of the smartphone 18 that notifies the results of the valve state prediction and diagnosis. In Figure 16 the example of the screen, it is shown that the current stage of the valve is the "stable" period, until the next "deterioration" stage, the remaining number of opening / closing operations is 25,830 and the predicted date and time of arrival is 2022 / 11 / 23, and the countermeasure policy to be adopted during the deterioration period is "The possibility of failure of important components increases, and it is a period that requires frequent inspections. It is recommended to perform inspections and replacements according to the importance of the valve."

[0072] 2.3. Summary of the Embodiment The valve state diagnosis device 2 according to the embodiment includes an angle sensor 4 that detects the operating angle of the valve and a processor 6. When the angle sensor 4 detects the operation of the valve, the processor 6 performs the following steps: (1) a step of obtaining an angle profile related to the operation of the valve from the angle sensor 4; (2) a step of histogramming the obtained angle profile, averaging the operations before and after, and performing relative frequency conversion; (3) a step of calculating the similarity between the histogram of the angle profile after relative frequency conversion and the histogram of the angle profile at the initial stage of operation by the histogram intersection method; (4) a step of plotting the calculated similarity of the angle profile in a two-dimensional plane where the similarity of the angle profile in the opening operation is taken on the horizontal axis and the similarity of the angle profile in the closing operation is taken on the vertical axis; and (5) a step of performing clustering processing on the plotted data of the similarity of the angle profile generated by repeatedly performing the plotting step in (4) above multiple times.

[0073] Such a valve state diagnosis device 2 can diagnose the state of the valve with a simple structure and represent in detail to the operator the change in the operation state of the valve and the change prediction.

[0074] 3. (Other embodiments) As described above, the embodiments have been described as examples of the technology disclosed in the present application. However, the technology in the present invention is not limited thereto, and can also be applied to embodiments that have been appropriately changed, replaced, added, omitted, etc.

[0075] The valve state diagnosis device 2 in the present invention can be applied not only to butterfly valves and ball valves, but also to multi-rotation on-off valves such as gate valves and globe valves.

[0076] In addition, as described above, the angle sensor 4 is not limited to a magnetic angle sensor. For example, it can also be an angle sensor using a potentiometer, an angle sensor using a rotary encoder, a gyro sensor, etc. Here, the angle sensor using a potentiometer is configured such that the output shaft of the air actuator of the on-off valve is connected to the shaft of the diagnosis unit equipped with the potentiometer, and this shaft rotates in accordance with the operation of the air actuator. The operation of this shaft is transmitted to the potentiometer via a gear, and a voltage corresponding to the value of the angle is output from the potentiometer. By measuring this voltage, the opening and closing angle of the valve is detected. In addition, the angle sensor using a rotary encoder is also configured such that the output shaft of the air actuator of the on-off valve is connected to the shaft of the diagnosis unit equipped with the rotary encoder, and this shaft rotates in accordance with the operation of the air actuator. The operation of this shaft is transmitted to the rotary encoder via a gear, and a number of pulses corresponding to the value of the angle is output from the rotary encoder. By counting these pulses, the opening and closing angle of the valve is detected. Further, the gyro sensor is a sensor that measures the angular velocity. Since it detects the "rotation angle per unit time", the rotation angle of the valve is obtained by integrating the detected angular velocity with respect to the time axis. In this way, the angle sensor 4 constituting the valve state diagnosis device 2 only needs to be able to detect or obtain the opening and closing angle of the valve in real time.

[0077] As described above, in the embodiment, for the following steps in the process of valve state diagnosis, it is performed at given intervals of the opening and closing operations of the valve. · Histogram similarity analysis (step S16). · Confirmation of the current status of the product life cycle (step S18). · Confirmation of changes in the product life cycle (step S20). · State prediction based on prior data (step S22). · State prediction based on the evaluation of the number of clusters (step S24). · State prediction / diagnosis result notification (step S26). For example, these steps can be performed at a given time interval (e.g., 1 month), or can be performed at given intervals of the opening and closing operations of the valve and at a given time interval. In addition, regarding the "state prediction / diagnosis result notification (step S26)", it can also be in the form of always sending information related to the latest valve state according to requests from the PC 16 and the smartphone 18.

[0078] In addition, for the purpose of illustrating the embodiments, drawings and detailed descriptions are provided. Therefore, among the components described in the drawings and the detailed descriptions, there are not only the components necessary to solve the technical problems, but also the components for exemplifying the above-mentioned technology but not necessary to solve the technical problems. Therefore, it should not be directly determined that these non-essential components are essential just because they are described in the drawings and the detailed descriptions.

[0079] In addition, since the above-described embodiments are used to illustrate the technology in the present invention, various changes, substitutions, additions, omissions, etc. can be made within the scope of the claims or their equivalents. Description of Reference Numerals

[0080] 2: Valve state diagnosis device, 4: Angle sensor, 6: Processor, 8: Memory, 10: Communication unit, 12: Cylinder, 14: Magnet, 16: PC (Personal computer), 18: Smart phone, 20: Cloud server, 21: Internet, 22: Valve unit, 24: USB connector, 26: Power supply and RS485 connector, 28: USB communication cable, 30: RS485 communication cable, 32: USB-RS485 converter, 34: Bluetooth.

Claims

1. A valve state diagnosis device, which is a valve state diagnosis device including an angle sensor and a processor, the angle sensor detecting the operating angle of the valve, when the angle sensor detects the operation of the valve, the processor performs the following steps: (1) The step of obtaining an angle profile related to the operation of the valve from the angle sensor; (2) The step of histogramming the obtained angle profile, averaging the operations before and after, and performing relative frequency conversion; (3) The step of calculating the similarity between the histogram of the relatively frequency-converted angle profile and the histogram of the angle profile at the initial stage of operation by the histogram intersection method; (4) The step of plotting the calculated similarity of the angle profile in a two-dimensional plane with the similarity of the angle profile in the opening operation on the horizontal axis and the similarity of the angle profile in the closing operation on the vertical axis; and (5) The step of performing clustering processing on the plotted data of the similarity of the angle profile generated by repeatedly performing the plotting step in (4) multiple times.

2. The valve state diagnosis device according to claim 1, wherein, the processor executes the plotting step in (4) and the step of performing clustering processing in (5) for the valve every given number of opening and closing operations.

3. The valve state diagnosis device according to claim 2, wherein, after the step of performing clustering processing in (5), the processor executes the step of inferring which period of the product life cycle the plotted data of the similarity of the latest angle profile corresponds to for the valve.

4. The valve state diagnosis device according to claim 2, wherein, after the step of performing clustering processing in (5), the processor executes the step of confirming the change in the product life cycle of the valve by confirming the increase in the number of clusters.

5. The valve state diagnosis device according to claim 3, wherein, after the step of performing clustering processing in (5) and the step of inferring which period of the product life cycle it corresponds to, the processor executes the step of grasping the position of the latest plotted data in the current product life cycle of the valve based on the existence ratio of each period of the product life cycle obtained in advance through experiments for each product type of the valve.

6. The valve state diagnosis device according to claim 2, wherein, in the step of performing clustering processing in (5), the processor executes the step of calculating the gap statistic for each assumed number of clusters for the valve, and predicting the progression of the product life cycle period based on the evaluation value obtained from the evaluation function in the gap statistic calculation and the progression of the difference between the evaluation values.

7. The valve state diagnosis device according to any one of claims 3 to 6, wherein, the valve state diagnosis device further includes a communication unit, the valve state diagnosis device sends the information generated by diagnosing the valve to an external computer via the communication unit.

8. The valve state diagnosis device according to claim 7, wherein, the angle sensor includes a magnetic angle sensor.

9. A valve state diagnosis method, including: (1) The step of obtaining an angle profile related to the operation of the valve by a processor; (2) A step of histogramming the obtained angular profile by the processor, and performing averaging of actions before and after to perform relative frequency conversion; (3) A step of calculating, by the processor, the similarity between the histogram of the relatively frequency-converted angular profile and the histogram of the angular profile at the initial stage of operation by the histogram intersection method; (4) A step of plotting, by the processor, the calculated similarity of the angular profile in a two-dimensional plane with the similarity of the angular profile in the opening action on the horizontal axis and the similarity of the angular profile in the closing action on the vertical axis; and (5) A step of performing clustering processing on the plotted data of the similarity of the angular profile generated by repeatedly performing the plotting step of (4) multiple times by the processor.

10. The valve state diagnosis method according to claim 9, wherein, For each given number of opening and closing operations of the valve, the processor executes the plotting step of (4) and the step of performing clustering processing of (5).

11. The valve state diagnosis method according to claim 10, wherein, After the step of performing clustering processing of (5), the processor executes a step of inferring which period of the product life cycle the plotted data of the similarity of the latest angular profile corresponds to.

12. The valve state diagnosis method according to claim 10, wherein, After the step of performing clustering processing of (5), the processor executes a step of confirming the change in the product life cycle of the valve by confirming the increase in the number of clusters.

13. The valve state diagnosis method according to claim 11, wherein, After the step of performing clustering processing of (5) and the step of inferring which period of the product life cycle it corresponds to, the processor executes a step of grasping the position of the latest plotted data in the current product life cycle of the valve based on the existence ratio of each period of the product life cycle obtained in advance by experiment for each product type of the valve.

14. The valve state diagnosis method according to claim 10, wherein, In the step of performing clustering processing of (5), the processor executes a step of calculating the gap statistic for each assumed number of clusters for the valve, and predicting the change in the period of the product life cycle of the valve based on the evaluation value obtained from the evaluation function in the gap statistic calculation and the trend of the difference between the evaluation values.

15. The valve state diagnosis method according to any one of claims 9 to 14, wherein, The valve state diagnosis method further includes a step of sending, by the processor, the information generated by diagnosing the valve to an external computer.

Images