Monitoring device, display device, monitoring method and monitoring program

Abstract

There are provided a monitoring device (10), a display device, a monitoring method, and a monitoring program capable of determining an operation state of a plant (100) even when a data amount of actually measured process data is limited. A monitoring device (10) includes: an input unit (10e, 11) that receives input of process data related to a plant (100); a model generation unit (12) that generates a model representing a relationship between the process data based on the input process data; a determination unit (13) that determines an operation state of the plant (100) in a first determination mode using the model generated based on a forecast value of the process data input by a person or in a second determination mode using the model generated based on an actually measured value of the process data; and a display unit (10f) that displays a determination result by the determination unit (13). The most suitable drawing: 1.

Description

Monitoring device, display device, monitoring method, and monitoring program

[0001] The present invention relates to a monitoring device, a display device, a monitoring method, and a monitoring program.

[0002] Conventionally, process data, which is time-series data relating to the operating state of a plant, is measured, and a model representing the relationship between the process data is generated by using the past process data as learning data. The generated model is sometimes used to determine whether the plant is operating normally.

[0003] For example, Patent Document 1 listed below describes a learning knowledge evaluation device that compares a target value when a control device controls a controlled object with past measured values ​​of the controlled object to determine whether the learning knowledge is appropriate.

[0004] Furthermore, the following Patent Document 2 describes a demand forecasting device that creates a forecasting model for predicting demand volume based on actual data on past demand volumes, and corrects the demand volume based on the actual data, future weather forecast data, and the predicted demand volume.

[0005] JP 7-219604 A JP 2018-73214 A

[0006] When generating a model that represents the relationship between process data by using past process data as learning data, it is assumed that a certain amount of actually measured process data has been accumulated. However, when the amount of actually measured process data is limited, such as immediately after starting up a plant, it becomes difficult to generate a model, and there may be a period during which the operating state of the plant cannot be determined.

[0007] Therefore, the present invention provides a monitoring device, a display device, a monitoring method, and a monitoring program that can determine the operating state of a plant even when the amount of actually measured process data is limited.

[0008] A monitoring device according to one aspect of the present invention includes an input unit that accepts input of process data related to a plant, a model generation unit that generates a model that represents a relationship between the process data based on the input process data, a determination unit that determines the operating state of the plant in a first determination mode that uses a model generated based on predicted values ​​of the process data input by a person or in a second determination mode that uses a model generated based on actual measured values ​​of the process data, and a display unit that displays a result of the determination made by the determination unit.

[0009] According to this aspect, by determining the operating state of the plant in the first determination mode using a model generated based on predicted values ​​of process data that have not been actually measured, even when the amount of actually measured process data is limited, it is possible to generate a model that represents the relationship between the process data and determine the operating state of the plant. This makes it possible to determine the operating state immediately after starting up the plant, thereby reducing downtime.

[0010] A display device according to another aspect of the present invention accepts input of process data related to a plant, generates a model representing the relationships between the process data based on the input process data, determines the operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a person, or in a second determination mode that uses a model generated based on actual measured values ​​of the process data, and displays the determination result.

[0011] A monitoring method according to another aspect of the present invention causes a monitoring device that monitors a plant to receive input of process data related to the plant, generate a model that represents a relationship between the process data based on the input process data, determine an operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a person or in a second determination mode that uses a model generated based on actual measured values ​​of the process data, and display a result of the determination.

[0012] A monitoring program according to another aspect of the present invention executes the following steps: accepting input of process data related to a plant; generating a model representing a relationship between the process data based on the input process data; determining the operating state of the plant in a first determination mode using the model generated based on predicted values ​​of the process data input by a human or in a second determination mode using the model generated based on actual measured values ​​of the process data; and displaying a determination result.

[0013] According to the present invention, it is possible to provide a monitoring device, a display device, a monitoring method, and a monitoring program that are capable of generating a model that represents the relationship between process data even when the amount of actually measured process data is limited.

[0014] FIG. 1 is a diagram showing functional blocks of a monitoring device according to an embodiment of the present invention; FIG. 2 is a diagram showing the physical configuration of a monitoring device according to an embodiment of the present invention; FIG. 3 is a diagram showing a model representing the relationship between process data, generated by a monitoring device according to an embodiment of the present invention; FIG. 4 is a diagram showing an abnormality degree calculated by a monitoring device according to an embodiment of the present invention; FIG. 5 is a flowchart of a determination process executed by a monitoring device according to an embodiment of the present invention; FIG. 6 is a diagram showing data points plotted by a monitoring device according to an embodiment of the present invention; and FIG. 7 is a flowchart of a model generation process executed by a monitoring device according to an embodiment of the present invention.

[0015] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described with reference to the accompanying drawings, in which the same reference numerals denote the same or similar components.

[0016] 1 is a diagram showing functional blocks of a monitoring device 10 according to an embodiment of the present invention. The monitoring device 10 is a device that monitors the operating state of a plant 100, and includes an acquisition unit 11, a model generation unit 12, a determination unit 13, a plot unit 14, an input unit 10e, and a display unit 10f.

[0017] The acquiring unit 11 acquires process data related to the plant 100. Here, the plant 100 may be any plant, but is targeted to plants from which process data can be acquired, such as power plants including boilers, incineration plants, chemical plants, and wastewater treatment plants. The process data may be any data related to the plant 100, but may be, for example, data obtained by measuring the state of the plant 100 with a sensor, and more specifically, may include measured values ​​such as the temperature, pressure, and flow rate of the plant 100. The acquiring unit 11 may acquire time-series data related to the plant 100 by acquiring the process data at predetermined time intervals or continuously.

[0018] The acquiring unit 11 may acquire multiple types of process data related to the plant 100. The acquiring unit 11 may acquire multiple types of process data measured by multiple sensors installed in the plant 100. Here, the multiple types of process data may be, for example, data representing different physical quantities such as temperature and pressure, or data representing the same physical quantity such as temperature measured at different locations in the plant 100.

[0019] The input unit 10e accepts input of process data. The input unit 10e may be configured with a touch panel, a pointing device such as a mouse, or a keyboard, and may accept input of predicted values ​​of process data predicted by a person. The input unit 10e may accept input of a handwritten graph to a drawing area where the process data is drawn, or may accept input of a handwritten range including the handwritten graph. The process data input by the input unit 10e will be described in detail later with reference to the drawings.

[0020] The display unit 10f displays a drawing area in which process data is drawn. The display unit 10f is used to monitor the operating state of the plant 100, and may display the determination result of the operating state of the plant 100 made by the monitoring device 10. The content displayed on the display unit 10f will be described in detail later with reference to the drawings.

[0021] The model generator 12 generates a model representing the relationship between the process data based on the input process data. The model representing the relationship between the process data may be a model representing a range into which the process data falls when the operating state of the plant 100 is normal, or a model extracting features that the process data exhibits when the operating state of the plant 100 is abnormal.

[0022] The model generator 12 may generate a model based on design values ​​of the process data. Here, the design values ​​of the process data refer to target values ​​or set values ​​at the time of plant design. More specifically, the design values ​​of the process data are values ​​of the process data that are to be measured in accordance with the design of the plant 100, and are values ​​of the process data that are measured when the plant 100 is operating normally. The model generator 12 may generate a model that represents the relationship between the process data by referring to the design values ​​of the process data that should be measured when the plant 100 is operating normally. By generating a model that represents the relationship between the process data based on the design values ​​of the process data, it is possible to generate a model that represents the relationship between the process data and determine the operating state of the plant 100 even when the amount of actually measured process data is limited.

[0023] The determination unit 13 determines the operating state of the plant 100 in a first determination mode using a model generated based on predicted values ​​of process data input by a human or a second determination mode using a model generated based on actual measured values ​​of the process data. Alternatively, the determination unit 13 may determine the operating state of the plant 100 using a model generated based on design values ​​of the process data. The model generated based on predicted values ​​of process data input by a human and the model generated based on actual measured values ​​of the process data may be the same model but using different data for model generation, or may be different models. Alternatively, the second determination mode may be a mode in which a model generated based on predicted values ​​of process data input by a human is updated based on actual measured values ​​of the process data. The model used in the second determination mode may be generated independently of the model used in the first determination mode, or may be a modified model of the model used in the first determination mode.

[0024] The monitoring device 10 according to the present embodiment determines the operating state of the plant 100 in the first determination mode using a model generated based on process data that has not been actually measured, so that even when the amount of actually measured process data is limited, it is possible to generate a model that represents the relationship between the process data and determine the operating state of the plant 100. This makes it possible to determine the operating state immediately after the plant 100 is started up, thereby shortening the start-up time when a new plant 100 is installed and reducing downtime when the plant 100 is temporarily stopped.

[0025] The determination unit 13 may calculate the degree of abnormality of the operating state of the plant 100 in the first determination mode or the second determination mode based on the actually measured process data, and determine the operating state of the plant 100 based on the degree of abnormality. Examples of the degree of abnormality calculated by the determination unit 13 will be described in detail later with reference to the drawings.

[0026] The plotting unit 14 plots data points representative of the handwritten graph received by the input unit 10e in the drawing area. The plotting unit 14 may also plot data points representative of the handwritten graph and handwritten range received by the input unit 10e in the drawing area. The processing by the plotting unit 14 will be described in detail later with reference to the drawings.

[0027] FIG. 2 is a diagram showing the physical configuration of a monitoring device 10 according to this embodiment. The monitoring device 10 includes a central processing unit (CPU) 10a, a random access memory (RAM) 10b, a read-only memory (ROM) 10c, a communication unit 10d, an input unit 10e, and a display unit 10f. These components are connected via a bus to enable data transmission and reception between them. While this example describes a case in which the monitoring device 10 is configured by a single computer, the monitoring device 10 may also be implemented by combining multiple computers. The configuration shown in FIG. 2 is merely an example, and the monitoring device 10 may include other components or may not include some of these components.

[0028] The CPU 10a is a control unit that controls the execution of programs stored in the RAM 10b or the ROM 10c and performs data calculations and processing. The CPU 10a is a calculation unit that generates a model representing the relationship between process data based on predicted values ​​of process data input by a human, and executes a program (monitoring program) that monitors the plant using the model. The CPU 10a receives various data from the input unit 10e and the communication unit 10d, displays the calculation results of the data on the display unit 10f, and stores them in the RAM 10b or the ROM 10c.

[0029] The RAM 10b is a rewritable memory device among the storage units, and may be configured, for example, with a semiconductor memory element. The RAM 10b may store data such as the program executed by the CPU 10a, process data input by a user, and design values ​​of the process data. Note that these are merely examples, and the RAM 10b may store data other than these, or may not store some of these data.

[0030] The ROM 10c is a memory from which data can be read, and may be configured, for example, with a semiconductor memory element. The ROM 10c may store, for example, a monitoring program and data that is not rewritten.

[0031] The communication unit 10d is an interface that connects the monitoring device 10 to other devices, and may be connected to a communication network N such as the Internet.

[0032] The input unit 10e receives data input from a user and may include, for example, a keyboard and a touch panel.

[0033] The display unit 10f visually displays the results of calculations performed by the CPU 10a and may be configured with, for example, an LCD (Liquid Crystal Display). The display unit 10f may display a drawing area in which process data is drawn, and may display the process data and the generated model in the drawing area.

[0034] The monitoring program may be provided by being stored in a computer-readable storage medium such as RAM 10b or ROM 10c, or may be provided via a communication network connected by communication unit 10d. In monitoring device 10, CPU 10a executes the monitoring program to realize acquisition unit 11, model generation unit 12, determination unit 13, and plot unit 14 described with reference to FIG. 1. Note that these physical configurations are merely examples and do not necessarily have to be independent configurations. For example, monitoring device 10 may include an LSI (Large-Scale Integration) in which CPU 10a is integrated with RAM 10b and ROM 10c.

[0035] 3 is a diagram showing a model representing the relationship between process data, which is generated by the monitoring device 10 according to this embodiment. The diagram shows an example in which predicted values ​​of process data are received from a person, and a model is generated based on the input predicted values ​​of the process data (process data that has not been actually measured).

[0036] The display unit 10f of the monitoring device 10 displays a drawing area DA in which process data is drawn. A user of the monitoring device 10 uses the input unit 10e to plot data points D1, which are predicted to represent the relationship between the first process data and the second process data, in the drawing area DA. The data points D1 may be input using a touch panel or a pointing device, or may be acquired from a storage unit built into the monitoring device 10, such as the RAM 10b, or from an external storage device. For example, the data points D1 may be design values ​​of the first process data and the second process data.

[0037] The model generation unit 12 generates a model representing the relationship between the first process data and the second process data based on the input data points D1. The model may include a graph M1 representing the relationship between the first process data and the second process data, and the display unit 10f may display the graph M1 in the drawing area DA. The model generation unit 12 may, for example, assume a predetermined function representing the relationship between the first process data and the second process data, and determine parameters of the function using the least squares method so that the function fits the data points D1. In this way, by displaying the graph M1 representing the relationship between the process data, it is possible to understand at a glance whether the generated model is appropriate.

[0038] The model may include a range M2 in which the process data falls near the graph M1 with a predetermined probability, and the display unit 10f may display the range M2 in the drawing area DA. The model generation unit 12 may, for example, calculate the standard deviation σ of the input data points D1 and set the range M2 to ±σ with the graph M1 at the center, or to ±2σ with the graph M1 at the center. When the variation in the process data follows a normal distribution, the range of ±σ is a range in which the process data falls near the graph M1 with a 68.27% probability, and the range of ±2σ is a range in which the process data falls near the graph M1 with a 95.45% probability. In this way, by displaying the range M2 in which the process data falls near the graph M1 with a predetermined probability, it is possible to grasp at a glance whether newly acquired process data is within the normal range.

[0039] 4 is a diagram showing the degree of abnormality calculated by the monitoring device 10 according to this embodiment. In the figure, the vertical axis indicates the value of the degree of abnormality, the horizontal axis indicates time, and the change in the degree of abnormality over time is shown in a bar graph.

[0040] The determination unit 13 of the monitoring device 10 may calculate the degree of abnormality of the operating state of the plant 100 in the first determination mode or the second determination mode based on the actually measured process data, and may display the degree of abnormality on the display unit 10f. The determination unit 13 may calculate the degree of abnormality using a known abnormality determination algorithm, for example, by using the mean μ and variance σ of the process data actually measured in the past. 2 Based on this, the abnormality degree a(x) of the current process data x is calculated as a(x) = (x - μ) 2 / σ 2In this case, the square root of the degree of abnormality represents how many times the standard deviation the current process data deviates from the average of the past process data. For example, if the degree of abnormality is 25, the current process data deviates from the average of the past process data by five times the standard deviation. The determination unit 13 may periodically calculate the degree of abnormality of the operating state of the plant 100 and display the value as a bar graph to produce the display shown in FIG. 4 , or may periodically calculate the degree of abnormality of the operating state of the plant 100 and display an average value over a longer period as a bar graph to produce the display shown in FIG. 4 . Furthermore, the determination unit 13 may calculate the degree of abnormality based on how much the actually measured process data deviates from the graph M1. For example, the determination unit 13 may calculate the degree of abnormality based on at least one of a value indicating whether the actually measured process data is inside or outside the range M2 and the amount of deviation of the actually measured process data from the standard deviation of the data point D1.

[0041] The determination unit 13 may determine the operating state of the plant 100 based on the calculated degree of abnormality. For example, the determination unit 13 may compare a threshold value set for the degree of abnormality with the newly calculated degree of abnormality, and determine that the operating state of the plant 100 is normal if the degree of abnormality is less than the threshold value, or determine that the operating state of the plant 100 is abnormal if the degree of abnormality is equal to or greater than the threshold value.

[0042] By displaying the abnormality degree as shown in FIG. 4, whether the operating state of the plant 100 is normal or abnormal can be expressed in quantitative numerical values, and even an operator who is not skilled in interpreting process data can make an appropriate judgment about the operating state of the plant 100.

[0043] 5 is a flowchart of the determination process executed by the monitoring device 10 according to this embodiment. First, the monitoring device 10 receives input of predicted values ​​of process data from a human (S10). Then, the monitoring device 10 generates a model representing the relationship between the process data based on the input predicted values ​​(S11). The model is used in the first determination mode.

[0044] The monitoring device 10 acquires the actually measured process data and stores it in the storage unit (S12). Then, the monitoring device 10 calculates the degree of anomaly of the plant 100 in a first determination mode using a model generated based on predicted values ​​of the process data input by a person, and determines the operating state of the plant 100 (S13). The monitoring device 10 displays the determination result on the display unit 10f (S14). Here, the monitoring device 10 may display the degree of anomaly and / or the actually measured process data together with the graph M1 and the range M2.

[0045] Thereafter, the monitoring device 10 determines whether the accumulated amount of the actually measured process data is equal to or greater than a predetermined amount (S15). Here, the predetermined amount may be an amount that allows a model representing the relationship between the process data to be generated based on the actually measured process data.

[0046] If the amount of accumulated measured process data is less than the predetermined amount (S15: NO), the monitoring device 10 acquires and accumulates new measured process data (S12), determines the operating state of the plant 100 in the first determination mode (S13), and displays the determination result (S14).

[0047] On the other hand, if the accumulated amount of actually measured process data is equal to or greater than the predetermined amount (S15: YES), the monitoring device 10 generates a model representing the relationship between the process data based on the actual measured values ​​of the accumulated process data (S16). Here, the monitoring device 10 may correct the model generated based on the predicted values ​​of process data that have not been actually measured in accordance with the actual measured values ​​of the process data, or may generate a new model using only the actual measured values ​​of the process data.

[0048] Thereafter, the monitoring device 10 acquires the actually measured process data and stores it in the storage unit (S17). Then, the monitoring device 10 calculates the degree of anomaly of the plant 100 in a second determination mode using a model generated based on the actual measured values ​​of the process data, determines the operating state of the plant 100 (S18), and displays the determination result (S19). In this case, the monitoring device 10 may also display the degree of anomaly and / or the actually measured process data together with the model. Note that, when a model generated based on a predicted value of process data input by a human and a model generated based on the actual measured values ​​of the process data are usable, the monitoring device 10 may accept a designation of which model to use. Furthermore, the monitoring device 10 may determine the operating state of the plant 100 based on the degree of anomaly calculated by a model generated based on the predicted value of process data input by a human and the degree of anomaly calculated by a model generated based on the actual measured values ​​of the process data.

[0049] 6 is a diagram showing data points plotted by the monitoring device 10 according to this embodiment. The diagram shows an example in which input of a handwritten graph G and a handwritten range R is received from a person, and data points D2 representing the input handwritten graph G and handwritten range R are plotted in the drawing area DA.

[0050] The display unit 10f of the monitoring device 10 displays a drawing area DA in which process data is drawn. A user of the monitoring device 10 uses the input unit 10e to input a graph G that is predicted as a relationship between the first process data and the second process data. The user also inputs a handwritten range R that includes the handwritten graph. Here, the handwritten range R may be a range in which the process data is predicted to fall within the vicinity of the handwritten graph G with a predetermined probability.

[0051] The plotting unit 14 of the monitoring device 10 plots data points D2 representing the handwritten graph G and the handwritten range R in the drawing area DA. The plotting unit 14 may, for example, plot the data points D2 so that they follow a normal distribution having a mean determined by the handwritten graph G and a variance determined by the handwritten range R, or may plot the data points D2 in the handwritten range R so that they follow a uniform distribution. After the data points D2 are plotted, the model generating unit 12 generates a model representing the relationship between the first process data and the second process data based on the data points D2.

[0052] By accepting input of a handwritten graph G and plotting data points D2 representing the handwritten graph G in the drawing area DA, the approximate relationships between the process data can be intuitively expressed by hand, and a model can be generated.

[0053] In addition, by accepting input of the handwritten range R and plotting the handwritten graph G and data points D2 representing the handwritten range R in the drawing area, the approximate relationship between the process data can be intuitively expressed by hand and a model can be generated.

[0054] 7 is a flowchart of a model generation process executed by the monitoring device 10 according to this embodiment. First, the monitoring device 10 receives input of a handwritten graph and a handwritten area from a person (S20). Then, the monitoring device 10 plots data points representing the handwritten graph and the handwritten area in a drawing area (S21).

[0055] Thereafter, the monitoring device 10 generates a model representing the relationship between the process data based on the data points (S22). Note that the monitoring device 10 may determine the operating state of the plant 100 in a first determination mode using the model thus generated.

[0056] The above-described embodiments are intended to facilitate understanding of the present invention and are not intended to limit the present invention. The elements of the embodiments, as well as their arrangement, materials, conditions, shapes, sizes, etc., are not limited to those illustrated and can be modified as appropriate. Furthermore, configurations shown in different embodiments can be partially substituted or combined with each other.

[0057] The display unit 10f of the monitoring device 10 is also a display device that receives input of process data related to the plant, generates a model representing the relationship between the process data based on the input process data, determines the operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a person, or in a second determination mode that uses a model generated based on actual measured values ​​of the process data, and displays the determination results. The display device may display at least one of the predicted values ​​of the process data input by a person, the generated model, and the actual measured values ​​of the process data together with the determination results.

[0058] REFERENCE SIGNS LIST 10...monitoring device, 10a...CPU, 10b...RAM, 10c...ROM, 10d...communication unit, 10e...input unit, 10f...display unit, 11...acquisition unit, 12...model generation unit, 13...determination unit, 14...plot unit, 100...plant

Claims

1. A monitoring device comprising: an input unit that accepts input of process data related to a plant; a model generation unit that generates a model that represents the relationship between the process data based on the input process data; a judgment unit that judges the operating state of the plant in a first judgment mode that uses the model generated based on predicted values ​​of the process data input by a human, or in a second judgment mode that uses the model generated based on actual measured values ​​of the process data; and a display unit that displays the judgment results made by the judgment unit.

2. The monitoring device according to claim 1, further comprising: the judgment unit calculates the degree of abnormality of the operating state of the plant in the first judgment mode or the second judgment mode based on the actually measured process data, and judges the operating state of the plant based on the degree of abnormality.

3. The monitoring device according to claim 1 or 2, wherein the model includes a graph representing a relationship between the process data, and the display unit displays the graph.

4. The monitoring device according to claim 3, wherein the model includes a range within which the process data falls near the graph with a predetermined probability, and the display unit displays the graph and the range.

5. The monitoring device according to any one of claims 1 to 4, wherein the model generation unit generates the model based on design values ​​of the process data.

6. A monitoring device as described in any one of claims 1 to 5, wherein the input unit accepts input of a handwritten graph into a drawing area in which the process data is drawn, and further comprises a plotting unit that plots data points representative of the handwritten graph in the drawing area, and the model generation unit generates the model based on the data points.

7. The monitoring device according to claim 6, wherein the input unit accepts input of a handwritten range including the handwritten graph in the drawing area, and the plot unit plots data points representing the handwritten graph and the handwritten range in the drawing area.

8. A display device that accepts input of process data related to a plant, generates a model that represents the relationships between the process data based on the input process data, determines the operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a person, or in a second determination mode that uses the model generated based on actual measured values ​​of the process data, and displays the determination results.

9. A monitoring method that causes a monitoring device that monitors a plant to perform the following steps: accept input of process data related to the plant; generate a model that represents the relationship between the process data based on the input process data; determine the operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a person, or in a second determination mode that uses the model generated based on actual measured values ​​of the process data; and display the determination results.

10. A monitoring program that causes a monitoring device that monitors a plant to execute the following steps: accept input of process data related to the plant; generate a model that represents the relationship between the process data based on the input process data; determine the operating state of the plant in a first determination mode that uses the model generated based on predicted values ​​of the process data input by a human or in a second determination mode that uses the model generated based on actual measured values ​​of the process data; and display the determination results obtained by the determination.