Abnormality diagnosis system, abnormality diagnosis method, and abnormality diagnosis program
The abnormality diagnosis system uses state change detection and scoring to identify the root cause of operational issues in complex systems by analyzing device interactions, optimizing the detection process for efficient fault localization.
Patent Information
- Application Number
- JP2021079824
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-05-10
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-05-10
AI Technical Summary
Existing abnormality diagnosis systems struggle to identify the specific device causing a problem when there are many devices in a facility or plant, especially when the abnormality affects multiple devices simultaneously.
An abnormality diagnosis system that uses state change detection means to identify changes in operational data, estimates the cause of these changes using inter-detection unit relationship information, and assigns scores based on evaluation criteria to determine the affected device, adjusting these scores to optimize the detection process.
The system effectively identifies the device causing an abnormal operating state by expressing the propagation relationship between devices, even in complex systems, facilitating easy identification of the root cause.
Smart Images

Figure 0007591970000001 
Figure 0007591970000002 
Figure 0007591970000003
Abstract
Description
[Technical field]
[0001] The present invention relates to an abnormality diagnosis system, an abnormality diagnosis method, and an abnormality diagnosis program that detect an abnormality during operation of a device, facility, plant, etc., and notify / transmit the abnormality to an operator or an operation system. [Background technology]
[0002] A method is known in which, in a facility or plant composed of multiple devices, operation data is obtained from measurement values obtained by sensors installed in each device, and abnormality diagnosis of the facility or plant is performed based on the operation data.
[0003] For example, Patent Document 1 describes an anomaly diagnosis method in which operation data is classified into a plurality of categories and recorded, and when operation data of a new category different from the data stored as operation data during normal operation is acquired, it is determined that the operation state is different from the operation state stored so far. Also, a method for determining an abnormality in equipment by considering the direction of material flow based on this method is described. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] Patent No. 4430384 Summary of the Invention [Problem to be solved by the invention]
[0005] However, in the technology disclosed in Patent Document 1, the devices in a facility or plant are diagnosed as a single group, so when there are many devices to be diagnosed or when the impact of an abnormality in one device is widespread, it is difficult to identify the device causing the problem.
[0006] An object of the present invention is to provide an abnormality diagnosis system, an abnormality diagnosis method, and an abnormality diagnosis program that can express the propagation relationship of influence between devices even when there are many devices in a facility or plant, and easily identify the device (part) that is causing an operating state that differs from normal operation. [Means for solving the problem]
[0007] In order to solve the above-mentioned problems, the present invention provides an abnormality diagnosis system for equipment in which a plurality of parts form a movement path for a material, the abnormality diagnosis system comprising a plurality of state change detection means, which are set for each part of the equipment and detect a change in state of the material, and a cause estimation means, which, when the state change detection means detects a change in state, estimates the part causing the change. The state change detection means stores a relationship during normal operation of a plurality of operation data related to the target part detected by the detection unit, and detects a state change when the relationship of the operation data has changed from the normal state. The cause estimation means assigns a score to the detection unit of each part that detects the state change using a plurality of evaluation criteria defined in relation to the material moving through the plurality of parts, and estimates the part causing the change in state of the equipment according to the total score assigned based on the plurality of evaluation criteria. The display device has a parameter adjustment means for adjusting the distribution of points for a plurality of evaluation criteria, and displays on the display device that the distribution adjustment of points for a plurality of evaluation criteria is performed by setting the distribution of points to be a linear function of the distance from the most upstream detection unit and the order of the units, and optimizing the intercept and slope of the linear function. The abnormality diagnosis system is characterized by:
[0008] The present invention also provides an abnormality diagnosis method for equipment in which a material movement path is formed by a plurality of parts, comprising: a plurality of state change detection means, which are set for each part of the equipment and detect a change in the state of the material, storing a relationship during normal operation of a plurality of operation data relating to the target part detected by the detection unit, and detecting a state change based on a change in the relationship of the operation data from a normal state; when the state change detection means detects a change in state, in order to estimate the part causing the change, a score is assigned to the detection unit of each part that detected the state change using a plurality of evaluation criteria defined in relation to the material moving through the plurality of parts; and the part causing the change in state of the equipment is estimated according to the total score assigned based on the plurality of evaluation criteria. The parameter adjustment means adjusts the distribution of scores for the multiple evaluation criteria, and the adjustment of the distribution of scores for the multiple evaluation criteria is set so that the distribution of scores is a linear function of the distance from the most upstream detection unit and the order of the units, and the display means displays that the intercept and slope of the linear function are optimized. The abnormality diagnosis method is characterized by:
[0009] The present invention also provides an abnormality diagnosis program for diagnosing abnormalities in equipment that forms a material movement path using a computer, the program comprising: a first program for storing the relationship of multiple operation data related to the target parts detected by the detection units during normal operation using multiple status change detection means that are set for each part of the equipment and detect a change in the status of the material, and detecting a status change based on a change in the relationship of the operation data from a normal state; a second program for assigning a score to the detection unit of each part that detected the status change using multiple evaluation criteria defined in relation to the material moving through the multiple parts in order to infer the part that is causing the change in status when the status change detection means detects a change in status, and inferring the part that is causing the change in status of the equipment based on the total score assigned based on the multiple evaluation criteria; and A third program to adjust allocations and a fourth program for displaying on a display means that the score distribution adjustment for the multiple evaluation criteria is set so that the score distribution is a linear function of the distance from the most upstream detection unit and the order of the units, and the intercept and slope of the linear function are optimized. The abnormality diagnosis program is characterized by comprising: Effect of the Invention
[0010] According to the present invention, even when there are many devices in a facility or plant, it is possible to express the propagation relationship of influence between the devices, and to easily identify the device (part) that is causing an operating state that differs from normal operation. [Brief description of the drawings]
[0011] [Figure 1] 1 is a block diagram showing an example of the configuration of an abnormality diagnosis system 1 according to an embodiment of the present invention. [Diagram 2] 1 is a diagram showing an example of the configuration of equipment in a plant that is a monitoring target of an abnormality diagnosis system 1. FIG. [Diagram 3] FIG. 13 is a diagram showing changes in operation data when a state change is detected in the detection unit A2. [Figure 4] FIG. 11 is a diagram showing an example of inter-detection unit relationship information. [Diagram 5] FIG. 2 is a diagram showing an example of a display means. [Figure 6]FIG. 13 is a diagram showing changes in operation data when state changes are detected in the detection units A2, A4, and A5. [Figure 7] 13 is a diagram showing an example of display of a control end and an operation end of a control logic. FIG. [Figure 8] FIG. 13 is a diagram showing the cause estimation results on the display means. [Figure 9] FIG. 13 is a diagram showing the cause estimation results on the display means in a different manner. [Figure 10] FIG. 2 is a diagram showing a processing flow of the abnormality diagnosis system 1 according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0012] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. EXAMPLES
[0013] 1 is a block diagram showing an example of the configuration of an abnormality diagnosis system 1 according to an embodiment of the present invention. The abnormality diagnosis system 1, as an example, is connected to a measurement and control device 3 of a plant 2, an external output device 4, and an external input device 5 so as to be capable of two-way communication. The abnormality diagnosis system 1, which is made up of a computer, also includes an operation data database 110, a state change detection means 120, an inter-detection unit relationship information database 130, a cause estimation means 140, a parameter adjustment means 150, and a display means 160.
[0014] Of these, the operation data database 110 is connected to the measurement and control device 3 of the plant 2, and stores measurement data such as the temperature, pressure, and flow rate of the plant, operation amount data such as valve opening, and control setting value data in chronological order. Hereinafter, the measurement data, operation amount data, and setting value data will be collectively referred to as operation data.
[0015] The status change detection means 120 detects status changes of each piece of equipment (which may be a part of the plant. Hereinafter, simply referred to as equipment) using a detection unit set for each piece of equipment in the plant. Hereinafter, the detection unit uses multiple pieces of operation data related to the target equipment and detects status changes from the relationship between these pieces of operation data.
[0016] The inter-detection unit relationship information database 130 stores the influence propagation relationship between the equipment (part) B of the plant corresponding to the detection unit A. For example, when the detection unit A1 and the detection unit A2 detect the state change of the equipment B1 and the equipment B2 of the plant, respectively, if the equipment B1 is an equipment on the upstream side of the equipment B2, the influence of the equipment B1 extends to the equipment B2. In this case, the inter-detection unit relationship information stores information that the influence of the detection unit A1 extends to the detection unit A2.
[0017] The cause estimation means 140 estimates the device B that caused the state change, using information on the state change detected by the detection unit A of the state change detection means 120 and the inter-detection unit relationship information database 130 .
[0018] The parameter adjustment means 150 adjusts the parameters used when the cause estimation means 140 estimates the device B that caused the state change.
[0019] The display means 160 displays the cause device B estimated by the cause estimation means 140 on the screen.
[0020] Next, the operation of the abnormality diagnosis system 1 will be described in detail, and as a prerequisite, the plant configuration example in Fig. 2 will be described as a reference example. The plant in Fig. 2 shows the transfer path of a substance through multiple devices B (in this example, devices B1 to B5), in which the substance from device B1 flows into device B2, which is mainly composed of a reactor V1, is diverted from device B2 to device B3 and device B4, which is mainly composed of a reactor V2, and further flows from device B4 to device B5. Control valves V- (V-F4, V-P5, V-F7) are appropriately arranged at each point of these paths to appropriately control the inflow and outflow of the substance.
[0021] In the operation data database 110 in Fig. 1, operation data for each part of the plant in Fig. 2 is detected by each sensor and recorded. As a specific example of the operation data here, for example, in the case of equipment B2 in the plant shown in Fig. 2, the flow rate F1, pressure P2, temperatures T1, T2, T3, and level L2 are the operation data used by the detection unit A2 corresponding to the equipment B2. Similarly, the operation data used by the detection unit A4 corresponding to the equipment B4 are the flow rate F4, pressure P6, temperatures T4, T6, and level L6.
[0022] Returning to Fig. 1, the anomaly diagnosis system 1 has a learning phase and a diagnosis phase. The learning phase is a preparation phase before the anomaly diagnosis system 1 is put into operation, in which only the state change detection means 120 operates. In the diagnosis phase, a diagnosis of device B of the plant is performed, and when the state change detection means 120 detects a change in the state of the device, the cause estimation means 140 operates. In addition, the diagnosis status in the diagnosis phase is displayed on the display means 160.
[0023] First, the learning phase will be described. In the learning phase, the operation data of the device B in a normal operating state stored in the operation data database 110 is used, and the detection unit A of the state change detection means 120 learns the relationship between the operation data of each part in a normal operating state by using, for example, the Adaptive Resonance Theory (hereinafter, referred to as ART), which is one of the data clustering techniques.
[0024] Specifically, multiple pieces of driving data corresponding to each detection unit are input to the ART as multidimensional data. The input here is the driving data shown in Figure 2.
[0025] The input operation data undergoes preprocessing such as normalization and addition of complements, and is then classified into multiple categories according to the similarity of the data. The number of categories to be classified varies depending on the resolution parameter that determines the level of detail of the classification, the number of operation data, the variance of the data, etc., but the classified categories are defined as categories that represent the normal state.
[0026] For example, if operational data obtained when a plant is in a normal state is divided into 10 categories, from category 1 to category 10, categories 1 to 10 are the categories for normal states.
[0027] Next, the diagnosis phase will be described. In the diagnosis phase, the operating data to be diagnosed (diagnosis data) is input to the ART, which has learned normal operating data. As a result, data that is highly similar to the learning data is classified into the same category as in the learning phase (one of 10 categories, 1 to 10). However, if some kind of abnormality occurs in the equipment and the trend of the data changes, it will be classified into a new category different from the learning data (a new category, for example, 11). In this way, the detection unit can detect changes in the state of the equipment from the categories classified by the ART.
[0028] Next, we will explain in detail what happens when a new category occurs in the diagnosis phase. Here, we will describe the case where a new category occurs in detection unit A2 and a state change is detected. In the case of the plant configuration in Figure 2, detection units A (A1 to A5) are provided in units of equipment B (B1 to B5), and detection unit A2 detects operation data (flow rate F1, pressure P2, temperatures T1, T2, T3, level L2) from sensors related to equipment B2, which is mainly composed of reactor V1 in Figure 2, and detects the abnormality by clustering.
[0029] The change in the values of each measurement item of detection unit A2 at this time is shown in Figure 3. The horizontal axis is the operating data item (flow rate F1, pressure P2, temperatures T1, T2, T3, level L2), and the vertical axis is the normalized measured value. The dashed line shows the characteristics of data in the same category as during normal operation (data immediately before a state change was detected), and the solid line shows the characteristics of data that became a new category. Detection unit A2 measured values for pressure P2 and temperature T3 that differed from those during normal operation, and this was the cause of the new category.
[0030] In addition, the operating data marked with "E" in Figure 3 indicates that it is an input condition for the target device B. For example, in the example of detection unit A2 in Figure 2, flow rate F1 and temperature T1 are data that depend on the upstream device B1 of the facility, and change regardless of the state of device B2. On the other hand, temperature T2, pressure P2, temperature T3, and level L2 are operating data that change depending on the upstream device B1, but are also affected by the state of reactor V1. In this way, the operating data that is the target of all detection units can be distinguished according to whether it is an input condition or not.
[0031] That is, the detection unit A can detect whether the state of the plant equipment B has changed, and at the same time, it can detect which operation data has changed from the characteristics of the classified data. It can also determine whether the changed operation data is an input condition.
[0032] Next, the operation of the cause estimation means 140 in Fig. 1 will be described. The cause estimation means 140 estimates the part that caused the abnormality from the time when each detection unit A detected a state change in the state change detection means 120, the operation data that caused the state change, and the propagation relationship of the influence between the parts of the equipment stored in the detection unit relationship information database 130.
[0033] FIG. 4 shows a specific example of information stored in the detection unit relationship information database 130. FIG. 4 shows the relationship between the detection units A1 to A5 corresponding to the devices B1 to B5 shown in FIG. 2. The arrows between the detection units indicate the propagation relationship of the influence. This is a propagation relationship determined by the movement of materials in the corresponding parts. For example, when focusing on the device B2 of the plant in FIG. 2, the material flowing in from the upstream device B1 moves to the devices B3 and B4 via the reactor V1. Therefore, in the detection unit A, the influence propagates from A1 to A2, and from A2 to A3 and A4 as shown in FIG. 4. In this way, the propagation relationship of the influence between the parts of the equipment stored in the detection unit relationship information database 130 represents the relationship between the detection units that collectively handle multiple operation data, and therefore can be constructed much more easily than a method of expressing the relationship between individual operation data.
[0034] The cause estimation means 140 estimates the part that caused the abnormality overall from the propagation relationship of these effects, the time when each detection unit A detected the state change in the state change detection means 120, and the operating data that caused the state change.
[0035] Fig. 10 is a flowchart showing the processing contents in the cause estimation means 140 and the parameter adjustment means 150. In Fig. 10, processing step S10 is processing in the parameter adjustment means 150 performed in a preparation stage before operating the cause estimation means 140, and adjusts parameters regarding standards used in the cause estimation means 140. Note that processing step S10 of the parameter adjustment means 150 will be described in detail later.
[0036] In process step S11, data is input and captured from each detection unit A (A1 to A5), and in process step S12, scores are assigned to the detection units A (A1 to A5) in accordance with the evaluation criteria shown below.
[0037] In this embodiment, the detection unit A that detected the state change was given a score based on the following four evaluation criteria, and the area with the highest total score was estimated to be the area most likely to be the cause of the abnormality. These evaluation criteria are defined in relation to a substance that moves through multiple areas.
[0038] Criterion 1: The detection unit that is further upstream in the influence propagation relationship is given a higher score.
[0039] Criterion 2: The detection unit with the faster detection time is given a higher score.
[0040] Criterion 3: Points are awarded to detection units whose driving data that caused the state change did not contain any input conditions.
[0041] Criterion 4: If the operational data that caused the state change is an operational end of the control logic, give points to the detection unit that includes the control end.
[0042] The validity of these criteria is judged as follows. First, criterion 1 is because the further upstream in the propagation relationship of effects, the more likely it is to be the cause of the anomaly. Criterion 2 is because the earlier a state change is detected, the more likely it is to be the cause. Criterion 3 is because operation data indicates that a state has changed even though the input conditions have not changed, and is therefore likely to be the root cause. Criterion 4 is because the root cause is likely to be on the control end side of the control logic. This is because operation data related to the control logic changes at the operation end but not at the control end.
[0043] Here, a specific example of criterion 4 will be explained with reference to Figure 2. In the plant equipment (site) B2, the level L2 of the reactor V1 is kept constant by the valve V-F4 that adjusts the flow rate F4 of the equipment (site) B4. At this time, there is a control logic between the level L2 and the flow rate F4 as shown by the dotted line, and the level L2 is called the control end and the flow rate F4 is called the operation end.
[0044] When such a control logic exists, the level L2 is kept constant, while the flow rate F4 fluctuates. Therefore, in criterion 4, when the operation data that caused the state change is the operation end of the control logic, it is considered that the root cause is likely to be the control end of the control logic.
[0045] Next, the specific method of awarding points for each criterion is as follows:
[0046] Criterion 1: The detection unit that is most upstream in the propagation relationship is awarded 100 points. The detection unit that is immediately downstream from the most upstream detection unit and that detects the state change is awarded 50 points. The detection unit that is second downstream from the most upstream detection unit and that detects the state change is awarded 30 points. The detection unit that is third downstream from the most upstream detection unit and that detects the state change is awarded 20 points. The detection unit that is fourth downstream from the most upstream detection unit and that detects the state change is awarded 10 points. No points are awarded to detection units that are fifth or more downstream from the most upstream detection unit.
[0047] Criterion 2: The unit with the fastest detection time is awarded 50 points. For each additional minute of delay in detection time, 1 point is deducted. For example, a unit that detects a change in status 15 minutes after the unit with the fastest detection time is awarded 35 points. Any unit that detects a change in status 50 minutes or more later is not awarded any points.
[0048] Criterion 3: 100 points are awarded to detection units whose operational data that caused the state change did not include the input condition.
[0049] Criterion 4: If the operational data that caused the state change is an operational end of a control logic, 50 points will be given to the detection unit that contains the control end of the corresponding control logic.
[0050] It should be noted that the method of awarding points in this embodiment is merely an example and does not limit the present invention.
[0051] In processing step S13 in Fig. 10, the sum of the points given to the detection units based on each criterion is calculated, and this content is displayed on the display means 160. An example of the display means 160 and an example of an actual cause inferred using these evaluation criteria are shown below. An example of the display of the display means 160 is shown in Fig. 5. The display means 160 shows a diagram representing the detection unit relationship information database 130. Fig. 5 shows information on the time when a state change was detected superimposed on Fig. 4, which shows the relationship between the detection units A1 to A5 corresponding to the devices B1 to B5.
[0052] In Figure 5, detection units A2, A4, and A5 with a colored background are detection units that were determined to have encountered a new category that differs from that during normal operation and detected a state change. The numbers in the upper right corner of detection units A2, A4, and A5 indicate the time at which the state change was detected. That is, detection unit A2 detected a state change at 8:30, detection unit A4 detected a state change at 8:20, and detection unit A5 detected a state change at 8:40.
[0053] In addition, in the example of Figure 5, the changes in the values of each measurement item of detection units A2, A4, and A5 are shown in Figure 6. The horizontal axis is the operating data item, and the vertical axis is the normalized measured value. The dashed line shows the characteristics of data in the same category as during normal operation (data immediately before a state change was detected), and the solid line shows the characteristics of data in a new category.
[0054] In Figure 6, "E" under the operation data indicates the input condition for the target device. In Figure 6, "C" under the operation data indicates that there is control logic and that it is a control end, and "O" indicates that there is control logic and that it is an operation end.
[0055] Another way of expressing the control logic is shown in Figure 7. In Figure 7, the control end and operation end of the control logic are summarized in a table. From Figure 7, it can be seen that the control end of control logic No. 1 is level L2, and the operation end is flow rate F4. Also, the control end of control logic No. 2 is level L6, and the operation end is flow rate F7.
[0056] According to the series of examples in Figures 5, 6, and 7, in detection unit A2, values different from those during normal operation were measured for pressure P2 and temperature T3, which was the reason for the new category. In detection unit A4, values different from those during normal operation were measured for flow rate F4 and temperature T4, which was the reason for determining the new category. In addition, the flow rate F4 and temperature T4 that caused the state change are both input conditions for detection unit A4. Furthermore, the flow rate F4 that caused the state change is the operation end of the control logic. In addition, in detection unit A5, values different from those during normal operation were measured for pressure P8 and temperature T8, which was the reason for the new category.
[0057] Based on these conditions, if we assign scores to each detection unit using the evaluation criteria shown above, we get the following:
[0058] Criterion 1: Detection unit A1 ~ 0 points, detection unit A2 ~ 100 points, detection unit A3 ~ 0 points, detection unit A4 ~ 50 points, detection unit A5 ~ 30 points Criterion 2: Detection unit A1 - 0 points, detection unit A2 - 40 points, detection unit A3 - 0 points, detection unit A4 - 50 points, detection unit A5 - 30 points Criterion 3: Detection unit A1 ~ 0 points, detection unit A2 ~ 100 points, detection unit A3 ~ 0 points, detection unit A4 ~ 0 points, detection unit A5 ~ 100 points Criterion 4: Detection unit A1 - 0 points, detection unit A2 - 50 points, detection unit A3 - 0 points, detection unit A4 - 0 points, detection unit A5 - 0 points Therefore, the total scores are: detection unit A1 - 0 points, detection unit A2 - 290 points, detection unit A3 - 0 points, detection unit A4 - 100 points, and detection unit A5 - 160 points. It can be estimated that the part causing the abnormality is detection unit A2, which has the highest total score.
[0059] In this embodiment, when the "Cause Estimation" button shown in Fig. 5 is pressed, the estimation result based on the assigned scores is displayed as shown in Fig. 8. Note that the "Probability" displayed in the rightmost column of Fig. 8 is evaluated based on the ratio of the score of each detection unit to the total score of the detection units.
[0060] In FIG. 5, when the cause estimation button is pressed, the screen that appears may be displayed in shades of color or hues as shown in FIG.
[0061] Next, the parameter adjustment means 150 will be described. The parameter adjustment means 150 adjusts the parameters (the score allocation method given by the criteria 1 to 4) used when the cause estimation means 140 estimates the device that caused the state change. This corresponds to the process of processing step S10 in FIG. 10.
[0062] The parameter adjustment is performed when it is determined that a difference has occurred between the result of the cause estimation means 140 estimating the equipment causing the state change and the equipment that would have caused the abnormality if it had occurred in the actual plant.
[0063] Specifically, the method of allocating points given in criteria 1 to 4 is optimized so that the total score of the equipment that caused an abnormality in an actual plant is maximized.
[0064] In criterion 1, for example, the score distribution is set to be a linear function of the distance from the most upstream detection unit and the order of the units, and the intercept and slope of the linear function are optimized.
[0065] In criterion 2, for example, the score distribution is set so that the detection time is a linear function of the elapsed time from the earliest detection unit, and the intercept and slope of the linear function are optimized.
[0066] Criterion 3, for example, optimizes the score given to a detection unit whose input condition is not included in the driving data that caused the state change.
[0067] In criterion 4, for example, when the operational data that caused the state change is an operation end of a control logic, the score given to the detection unit including the control end of the corresponding control logic is optimized.
[0068] By optimizing the point allocation method given by criteria 1 to 4 (adjusting the parameters), it is possible to match the equipment (part) that caused an abnormality when it occurs in an actual plant with the result of the cause estimation means 140 estimating the equipment (part) that caused the state change.
[0069] As described above, in this embodiment, a change in the state of a device (part) of a plant can be detected, and the device (part) that is the cause of the state change can be estimated.
[0070] In this embodiment, the adaptive resonance theory is used as the clustering technique for the state change detection means 120, but other data clustering techniques such as vector quantization and k-means may also be used.
[0071] In the above explanation, the explanation has been given mainly from the viewpoint of the system configuration, but this can be realized as a method or a program. When realized as a method, it is preferable to use "an abnormality diagnosis method for equipment in which a movement path of a material is formed by a plurality of parts, characterized in that a plurality of state change detection means, which are set for each part of the equipment and detect a change in state of the material, store the relationship of a plurality of operation data related to the target part detected by the detection unit during normal operation, detect a state change based on a change in the relationship of the operation data from the normal state, and when the state change detection means detects a change in state, in order to estimate the part causing the change, a score is assigned to the detection unit of each part that detected the state change using a plurality of evaluation criteria defined in relation to the material moving through the plurality of parts, and the part causing the change in state of the equipment is estimated according to the total score assigned by the plurality of evaluation criteria."
[0072] Furthermore, when it is realized as a program, it is preferable to define it as "an abnormality diagnosis program that uses a computer to diagnose abnormalities in equipment that forms a movement path for a material using multiple parts, comprising: a first program that uses multiple state change detection means that are set for each part of the equipment and detect changes in the state of the material to store the relationship between multiple operating data for the target part detected by the detection unit during normal operation and detects a state change when the relationship between the operating data has changed to a normal state; a second program that, when the state change detection means detects a change in state, assigns points to the detection unit of each part that detected the state change using multiple evaluation criteria defined in relation to the material moving through the multiple parts in order to infer the part that is causing the change in state, and infers the part that is causing the change in state of the equipment depending on the total score assigned based on the multiple evaluation criteria; and a third program that adjusts the distribution of points based on the multiple evaluation criteria." [Explanation of symbols]
[0073] 1. Abnormality diagnosis system 2. Plant 3…Measurement and control equipment 4. External output device 5…External input device 110…Operation data database 120...State change detection means 130…Detection unit relationship information database 140...Cause estimation means 150...Parameter adjustment means 160...Display means
Claims
1. An abnormality diagnosis system for equipment that forms a material movement path by a plurality of parts, The system comprises a plurality of state change detection means, which are set for each part of the equipment and detect a state change of a substance, and a cause estimation means, which, when the state change detection means detects a state change, estimates a part causing the change, The state change detection means stores a relationship between a plurality of pieces of driving data related to the target part detected by the detection unit during normal driving, and detects a state change when the relationship between the driving data has changed to a normal state; the cause estimation means assigns a score to the detection unit of each part that detects a change in state by using a plurality of evaluation criteria defined in relation to a substance moving through a plurality of parts, and estimates the part that causes the change in state of the equipment according to the total score assigned based on the plurality of evaluation criteria; A parameter adjustment means for adjusting the distribution of points in a plurality of evaluation criteria, An abnormality diagnosis system characterized in that the adjustment of the score distribution for multiple evaluation criteria is set so that the score distribution is a linear function of the distance from the most upstream detection unit and the order of the units, and the display means displays that the intercept and slope of the linear function are optimized.
2. An abnormality diagnosis system for equipment that forms a material movement path by multiple parts, comprising: The system comprises a plurality of state change detection means, which are set for each part of the equipment and detect a state change of a substance, and a cause estimation means, which, when the state change detection means detects a state change, estimates a part causing the change, The state change detection means stores a relationship between a plurality of pieces of driving data related to the target part detected by the detection unit during normal driving, and detects a state change when the relationship between the driving data has changed to a normal state; the cause estimation means assigns a score to the detection unit of each part that detects a change in state by using a plurality of evaluation criteria defined in relation to a substance moving through a plurality of parts, and estimates the part that causes the change in state of the equipment according to the total score assigned based on the plurality of evaluation criteria; A parameter adjustment means for adjusting the distribution of points in a plurality of evaluation criteria, An abnormality diagnosis system characterized in that the adjustment of the score distribution for multiple evaluation criteria is performed by setting the score distribution so that it is a linear function of the elapsed time from the detection unit with the earliest detection time, and by displaying on a display means that the intercept and slope of the linear function are optimized.
3. An abnormality diagnosis system for equipment that forms a material movement path by multiple parts, comprising: The system comprises a plurality of state change detection means, which are set for each part of the equipment and detect a state change of a substance, and a cause estimation means, which, when the state change detection means detects a state change, estimates a part causing the change, The state change detection means stores a relationship between a plurality of pieces of driving data related to the target part detected by the detection unit during normal driving, and detects a state change when the relationship between the driving data has changed to a normal state; the cause estimation means assigns a score to the detection unit of each part that detects a change in state by using a plurality of evaluation criteria defined in relation to a substance moving through a plurality of parts, and estimates the part that causes the change in state of the equipment according to the total score assigned based on the plurality of evaluation criteria; A parameter adjustment means for adjusting the distribution of points in a plurality of evaluation criteria, An abnormality diagnosis system characterized by displaying on a display means that the adjustment of the distribution of scores for multiple evaluation criteria is intended to optimize the scores given to detection units whose input conditions are not included in the operating data that caused the state change.
4. An abnormality diagnosis system for equipment that forms a material movement path by multiple parts, comprising: The system comprises a plurality of state change detection means, which are set for each part of the equipment and detect a state change of a substance, and a cause estimation means, which, when the state change detection means detects a state change, estimates a part causing the change, The state change detection means stores a relationship between a plurality of pieces of driving data related to the target part detected by the detection unit during normal driving, and detects a state change when the relationship between the driving data has changed to a normal state; the cause estimation means assigns a score to the detection unit of each part that detects a change in state by using a plurality of evaluation criteria defined in relation to a substance moving through a plurality of parts, and estimates the part that causes the change in state of the equipment according to the total score assigned based on the plurality of evaluation criteria; A parameter adjustment means for adjusting the distribution of points in a plurality of evaluation criteria, An abnormality diagnosis system characterized in that the adjustment of the distribution of scores for multiple evaluation criteria is displayed on a display means to optimize the score to be given to a detection unit including the control end of the corresponding control logic when the operating data that caused the state change is the operating end of a control logic.
5. The abnormality diagnosis system according to claim 1, An anomaly diagnosis system characterized in that one of the multiple evaluation criteria is to give a higher score to a detection unit that is located upstream in the propagation relationship of the influence of a substance moving through multiple parts.
6. The abnormality diagnosis system according to claim 2, An anomaly diagnosis system characterized in that one of the multiple evaluation criteria is to give a higher score to detection units that have a faster detection time for a substance moving through multiple locations.
7. The abnormality diagnosis system according to claim 3, An anomaly diagnosis system characterized in that one of the multiple evaluation criteria is to give points to detection units whose input conditions are not included in the operating data that caused the state change for a substance moving through multiple parts.
8. The abnormality diagnosis system according to claim 4, An abnormality diagnosis system characterized in that one of the multiple evaluation criteria is to give a score to a detection unit including a control end when the operating data that caused a change in state for a substance moving through multiple parts is the operating end of a control logic.
9. A method for diagnosing an abnormality in equipment that forms a material movement path by a plurality of parts, comprising: A plurality of state change detection means are set for each part of the equipment and detect a state change of a substance, storing a relationship of a plurality of operation data related to the target part detected by the detection unit during normal operation, and detecting a state change when the relationship of the operation data has changed from the normal state; When the status change detection means detects a status change, in order to estimate the part causing the change, a score is assigned to the detection unit of each part that detects the status change using a plurality of evaluation criteria defined in relation to a substance moving through a plurality of parts, and the part causing the status change of the equipment is estimated according to the total score assigned based on the plurality of evaluation criteria; A parameter adjustment means adjusts the distribution of scores among a plurality of evaluation criteria; An abnormality diagnosis method characterized in that the adjustment of the score distribution for multiple evaluation criteria is performed by setting the score distribution to be a linear function of the distance from the most upstream detection unit and the order of the units, and displaying on a display means that the intercept and slope of the linear function are optimized.
10. An abnormality diagnosis program for diagnosing abnormalities in equipment that forms a material movement path using a plurality of parts, comprising: a first program for storing a relationship during normal operation of a plurality of operation data related to a target part detected by a detection unit using a plurality of state change detection means set for each part of the equipment and detecting a state change of a substance, and detecting a state change based on a change in the relationship of the operation data from the normal state; a second program for assigning points to the detection units of each part that detects a change in status using multiple evaluation criteria defined in relation to a substance moving through multiple parts in order to infer the part that causes the change in status when the status change detection means detects a change in status, and inferring the part that causes the change in status of the equipment according to the total score assigned based on the multiple evaluation criteria; and a third program for adjusting the distribution of points based on the multiple evaluation criteria. An abnormality diagnosis program characterized by having a fourth program that displays on a display means that the adjustment of the score distribution for multiple evaluation criteria is performed by setting the score distribution to be a linear function of the distance from the most upstream detection unit and the order of the units, and optimizing the intercept and slope of the linear function.
Citation Information
Patent Citations
Method and system for elimination of fault conditions in a technical installation
EP3660612A1
Failure diagnosis device
JP1992360206A
Equipment diagnostic device and diagnostic method
JP4430384B2
Abnormality diagnosis system
WO2017159016A1
