Systems and methods for supporting manufacturing

Abstract

To provide information which serves as a guideline for determination and operation by a user who is in charge of operation.SOLUTION: A system for supporting production includes: a plurality of sensors for measuring the amount of state and the amount of operation of a production facility; a data collection / management unit which collects and holds time-series data of multiple variables measured by the sensors; a variable definition unit which defines a variable to be analyzed, from the held data and variables measured online; a data forming processing unit and a principal component analysis processing unit which pre-process the variables and perform principal component computation; a variable correlation display unit which executes multivariate analysis for each predetermined cycle, updates analysis values, displays them as time-series information, and displays the time-series information serving as a target index, in conformity with time-series principal component values; a state change definition unit which defines a state change to be subjected to factor variable extraction from multi-time series information; a change factor variable calculation processing unit which executes a factor variable extraction process with a high correlation on the defined state change; and a change factor variable display unit which displays a factor variable related to the extracted state change.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a system and method for providing guidance and support to users who operate manufacturing equipment and machinery when adjusting the equipment and machinery conditions. [Background technology]

[0002] Patent Document 1 discloses a technology that processes multidimensional time-series data obtained based on detection signals output by numerous sensors placed in manufacturing equipment and machinery (hereinafter simply referred to as "manufacturing equipment") using principal component analysis, and notifies the status of the manufacturing equipment in a visually easy-to-understand output format. Furthermore, a predictive diagnostic system that detects signs of abnormalities based on the analysis of multidimensional time-series data is known, for example, from Patent Document 2.

[0003] These systems utilize multivariate analysis, such as principal component analysis, to diagnose conditions by combining information from numerous sensors. By comprehensively and holistically monitoring the state of manufacturing equipment, they can detect early signs of malfunctions and notify users. This reduces the proportion of corrective maintenance and provides advanced support for the operation of manufacturing equipment. Furthermore, principal components that influence output variables are extracted, and a large amount of normal data is collected and used for learning, and statistical measures are used to determine thresholds for abnormality and normality in order to predict manufacturing equipment malfunctions.

[0004] Furthermore, a system for displaying factor variables strongly related to the analysis results of principal component analysis is known from Patent Document 3. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2001-67117 [Patent Document 2] Japanese Patent Publication No. 2019-204342 [Patent Document 3] Japanese Patent Publication No. 2014-178844 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] For example, in manufacturing facilities such as chemical plants, productivity can fluctuate during operation, even if it's not a malfunction. Productivity refers to factors such as unit consumption, yield rate, and energy efficiency of manufacturing. These fluctuations are due to various factors, including environmental changes such as temperature, changes in the quality of catalysts and other components in the manufacturing process, and the effects of maintenance. While stopping production for maintenance is one option when productivity declines, completely halting production has disadvantages, such as a significant decrease in output volume and increased maintenance costs. Therefore, continuing production is also an option, but it is desirable to improve productivity, even slightly.

[0007] As described in Patent Documents 1 to 3 above, many technologies for predicting and identifying normal and abnormal conditions have been known for some time. However, to improve productivity even slightly within the range of normal conditions, there is still a reliance on the experience of skilled technicians.

[0008] However, considering the retirement of skilled technicians due to aging and the anticipated shortage of personnel to handle operations in the future due to the declining birthrate, it is desirable to formalize the productivity improvement knowledge possessed by skilled technicians, which will ultimately lead to automation.

[0009] Therefore, the object of the present invention is to provide a manufacturing support system and a manufacturing support method that extract factors in response to various changes in operating conditions while the manufacturing equipment is operating within a normal range, and provide information that serves as a guideline to support the judgment and operation of the user performing the operation. [Means for solving the problem]

[0010] To solve the above problems, the present invention is characterized by comprising: a plurality of sensors for measuring state quantities or manipulated quantities of manufacturing equipment; a data acquisition and management unit for collecting and storing time-series data of a plurality of variables measured by the plurality of sensors; a variable definition unit for defining variables to be analyzed from the stored data and online measured variables; a data shaping processing unit and a principal component analysis processing unit for pre-processing the variables defined in the variable definition unit and performing principal component calculations; a variable correlation display unit for performing multivariate analysis according to data for a predetermined period at predetermined cycles, updating analysis values ​​based on the multivariate analysis for a predetermined period at predetermined cycles and displaying them as time-series information, and displaying time-series information that serves as a target indicator in accordance with the time-series principal component values; a state change definition unit for defining state changes that are the target of factor variable extraction from the time-series information based on the multivariate analysis; a change factor variable calculation processing unit for performing factor variable extraction processing that has a large correlation with the state changes defined by the state change definition unit; and a change factor variable display unit for displaying factor variables related to state changes extracted by the factor variable extraction processing.

[0011] Furthermore, the present invention is A method for supporting manufacturing using a manufacturing support system, comprising: (a) a step in which multiple sensors measure state quantities or operation quantities of manufacturing equipment; (b) a step in which a data acquisition and management unit collects and stores time-series data of multiple variables measured by the multiple sensors; (c) a step in which a variable definition unit defines variables to be analyzed from the stored data and online measured variables; (d) a step in which a data formatting processing unit and a principal component analysis processing unit preprocess the variables defined in the variable definition unit and perform principal component calculations; and (e) a variable correlation display unit, The steps include: performing multivariate analysis according to data for a predetermined period at predetermined cycles; updating the analysis values ​​based on the multivariate analysis for the predetermined period at predetermined cycles and displaying them as time-series information; and displaying the time-series information that serves as the target indicator in accordance with the time-series principal component values. and , (f) The state change definition unit, The step of defining state changes that are the target of factor variable extraction from the time series information based on the multivariate analysis. and , (g) The variable calculation processing unit for the change factor, The aforementioned In the state change definition section Steps to perform a process to extract factor variables that have a high correlation with the defined state changes. and , (h) The variable display section for the change factor is, The aforementioned By extracting factor variables Step to display the factor variables related to the extracted state changes. and It is characterized by including , [Effects of the Invention]

[0012] According to the present invention, it is possible to realize a system for supporting manufacturing that extracts factors corresponding to various changing operating conditions while the manufacturing equipment is operating within a normal range, and provides information serving as a guideline for assisting the judgment and operation of a user who performs operations, and a method for supporting manufacturing.

[0013] Problems, configurations, and effects other than those described above will be clarified by the description of the embodiments for carrying out the following invention.

Brief Description of the Drawings

[0014] [Figure 1] It is a diagram showing an example of time-series data of the productivity of manufacturing equipment. [Figure 2] It is a block diagram showing a schematic configuration of a chemical plant management system according to Embodiment 1 of the present invention. [Figure 3] It is a diagram showing an example of manufacturing equipment 200 and a sensor group 300. [Figure 4] It is a functional block diagram showing the configuration and processing flow of a plant monitoring and control support system 100. [Figure 5A] It is a diagram showing an example of a display of a state change display unit 108. [Figure 5B] It is a diagram showing an example of a display of a state change display unit 108. [Figure 6A] It is a diagram showing an example of a display of a state change display unit 108. [Figure 6B] It is a diagram showing an example of a display of a state change display unit 108. [Figure 7] It is a diagram showing an example of an operation in a state change definition unit 110. [Figure 8] It is a diagram showing an example of displaying the inner product values of all variables in the order of processes. [Figure 9] It is a diagram showing an example of a display in the order of the magnitude of the inner product. [Figure 10] It is a diagram showing an example of data of each part of a chemical plant management system according to Embodiment 2 of the present invention. [Figure 11] It is a diagram showing an example of a prognostic diagnosis system and an abnormality detection system 600. [Figure 12]This figure shows the change in sensor values. [Figure 13] This diagram conceptually illustrates the differences in user behavior expected depending on whether or not the present invention is applied. [Modes for carrying out the invention]

[0015] Embodiments of the present invention will be described below with reference to the drawings. The embodiments are illustrative examples for explaining the present invention, and have been omitted and simplified as appropriate for clarity of explanation. The present invention can also be carried out in various other forms. Unless otherwise specified, each component may be singular or plural.

[0016] Furthermore, the position, size, shape, and range of each component shown in the drawings may not represent the actual position, size, shape, and range in order to facilitate understanding of the invention. For this reason, the present invention is not necessarily limited to the position, size, shape, and range disclosed in the drawings. [Examples]

[0017] Referring to Figures 1 to 9, a system and method for supporting manufacturing according to Embodiment 1 of the present invention will be described.

[0018] First, the basic concept of the present invention will be explained using Figure 1. Figure 1 shows an example of time-series data of the productivity of manufacturing equipment. The horizontal axis represents time, and the vertical axis represents productivity indicators such as the operating rate of the manufacturing equipment and the yield rate of good products.

[0019] In Figure 1, productivity refers to indicators such as the operating rate of manufacturing equipment, the yield rate of good products, and production volume. Anomaly detection, on the other hand, refers to detecting serious problems such as equipment failure or shutdown, or a significant decrease in productivity.

[0020] For example, in manufacturing facilities such as chemical plants, annual production plans are sometimes formulated based on a baseline productivity. If productivity falls below the plan and there is room for improvement, B: an attempt is made to continue operations while suppressing the decline in productivity. In this case, there is a possibility that it may lead to sudden failures or shutdowns of manufacturing equipment in the future, or unexpected declines in product quality.

[0021] On the other hand, C: If an abnormality or an indication of an abnormality is detected and production is stopped, and the cause of the abnormality is removed and production is restarted, stopping the equipment for maintenance will sacrifice the operating rate and, consequently, the overall productivity.

[0022] Therefore, in this invention, A: by continuing trials to improve production efficiency within a normal range, it contributes to improving the manufacturing capacity of the target equipment itself. To this end, it extracts factors in response to various changes in operating conditions while the manufacturing equipment is operating within a normal range, and provides information that serves as a guideline to support the judgment and operation of the user performing the operation.

[0023] The system supporting manufacturing in this embodiment will be described using Figure 2. Figure 2 is a block diagram showing the schematic configuration of the chemical plant management system 1, including the plant monitoring and control support system 100 of this embodiment.

[0024] As shown in Figure 2, the plant monitoring and control support system 100 of this embodiment monitors the operation of the manufacturing equipment 200 based on detection signals from a sensor group 300 consisting of numerous sensors, extracts slight state changes and their contributing factors, and supports the efficient operation of the manufacturing equipment 200 by the user.

[0025] The chemical plant management system 1 shown in Figure 2, as an example, includes a productivity data management system 400, a manufacturing management system 500, a predictive diagnostic system and anomaly detection system 600, a monitoring and control unit 700, and a data acquisition and management system 800.

[0026] Although the specification describes the productivity data management system 400, manufacturing management system 500, predictive diagnostic system and anomaly detection system 600, monitoring and control unit 700, and data acquisition and management system 800 separately as functions, since the present invention relates to the plant monitoring and control support system 100, all or some of these may be configured within the same device as the plant monitoring and control support system 100.

[0027] Various data based on detection signals output from the sensor group 300 are input to the monitoring and control unit 700 and transmitted through the data acquisition and management system 800 to the productivity data management system 400, the predictive diagnostic system and anomaly detection system 600, and the plant monitoring and control support system 100.

[0028] The plant monitoring and control support system 100 diagnoses the status of the manufacturing equipment 200 and fluctuations in the manufacturing environment, and displays and outputs information that serves as a guideline for parameter control of the manufacturing equipment 200. Its detailed functions and structure will be described later.

[0029] The Productivity Data Management System 400 manages productivity data of products measured and inspected, for example, hourly, daily, per raw material lot, or per manufacturing lot. Productivity data includes, for example, the yield rate of products, the unit cost (raw material consumed to produce a certain amount of product), and the energy efficiency per unit of production during the manufacturing process.

[0030] To give a specific example, the productivity data management system 400 has the function of importing and managing analysis results from an analytical device (not shown) that analyzes the quality of raw materials or finished products. In addition, the productivity data management system 400 receives data on raw material lots and manufacturing lots from the manufacturing management system 500, and manages productivity data according to this data on a period-by-period basis, or on a raw material lot-by-lot basis or manufacturing lot-by-lot basis.

[0031] The manufacturing management system 500 is a system that manages raw materials used in manufacturing at the manufacturing equipment 200, raw material lots and manufacturing lots, inventory, finished products, and information related to the control of the manufacturing equipment 200, and manages manufacturing by the manufacturing equipment 200 through the monitoring and control unit 700.

[0032] The monitoring and control unit 700 has the function of monitoring detection signals from the sensor group 300 and controlling the manufacturing equipment 200 based on that data. In addition, the monitoring and control unit 700 controls the manufacturing equipment 200 in accordance with various information managed by the manufacturing management system 500.

[0033] Figure 3 shows an example of the manufacturing equipment 200 and sensor group 300. The manufacturing equipment 200, as an example, includes a raw material tank 201, a storage tank 202, a reaction tank 203, a reaction tank 204, and a product storage tank 205, and is a manufacturing facility for producing the final product C from raw material groups S and M.

[0034] Tanks 201-205 are connected by piping, through which raw materials, intermediate products, and final products are transported. In addition to raw materials, piping is also connected to supply nitrogen and oxygen. Furthermore, numerous heaters, valves, etc. (not shown) are installed independently or in combination with sensors, and adjusting these also causes the detected values ​​of sensor group 300 to change. In reaction tanks 203 and 204, the final product C is produced by reacting raw material groups S and M and gas through stirring, heating, pressurization, etc.

[0035] The sensor group 300 includes, as an example, a flow sensor for measuring the flow rate of raw materials, a temperature sensor for measuring the temperature of the reaction tank 203, a pressure sensor for measuring the pressure of the reaction tank 204, a level sensor for measuring the amount of product stored in the product storage tank 205, and a concentration sensor for measuring the concentration of the product.

[0036] Furthermore, the sensor group 300 also includes flow sensors that measure the supply rate of gases such as nitrogen and oxygen, and it goes without saying that the types of sensors in the sensor group are not limited to these, and their arrangement is just one example. For example, temperature sensors can be installed not only in the reaction tank 203 but also in multiple locations within the manufacturing equipment 200. In addition, the objects measured by the temperature sensors may include not only raw materials, intermediate products, and final products, but also oil, cooling water, and ambient temperature.

[0037] In such a manufacturing facility 200, each of the multiple sensors in the sensor group 300 outputs a detection signal, and the monitoring and control unit 700 detects whether the value of each detection signal is a normal value that falls within a predetermined range (upper limit, lower limit) or an abnormal value that exceeds this range.

[0038] Furthermore, each detected value from the sensor group 300, including data not directly related to control, is collected and stored by the data acquisition and management system 800, and becomes a database for the productivity data management system 400, the predictive diagnostic system and anomaly detection system 600, and the plant monitoring and control support system 100. The data acquisition and management system 800 may also have data formatting functions, such as unifying the timestamp information of each detected value from the sensor group, in addition to data storage.

[0039] The output values ​​of the predictive diagnostic system and anomaly detection system 600 are displayed in the same way as the detection signals of the sensor group 300, allowing for the display of the degree of anomaly obtained by comprehensively monitoring the manufacturing equipment 200.

[0040] In this invention, the plant monitoring and control support system 100 further emphasizes and displays relatively small state changes. While it is similar to the process monitoring and diagnostic device in Patent Document 3 in that it performs principal component analysis on multidimensional time-series data obtained from the sensor group 300, this system does not distinguish between normal and abnormal, but rather monitors and displays information on state changes based on principal component analysis to support control.

[0041] Therefore, this invention does not have a learning model. Since it is not a learning-based system, no data for learning is required. Not relying on learning has the advantage of being able to respond flexibly to changes in the state of manufacturing equipment, such as changes in climate, and to analyze based on the most recent data. However, this does not preclude the use of this invention in combination with a monitoring system that has a learning model. For example, a predictive diagnostic system or anomaly detection system that has a learning model can be incorporated as shown in Figure 2.

[0042] Figure 4 shows the configuration and processing flow of the plant monitoring and control support system 100. To perform the above processing, the plant monitoring and control support system 100 is composed of, for example, an offline data storage unit 101, an online data acquisition unit 102, a variable definition unit 103, a data formatting processing unit 104, a principal component analysis processing unit 105, an indicator definition unit 106, a clustering processing unit 107, a state change display unit 108, a variable correlation display unit 109, a state change definition unit 110, a change factor variable calculation processing unit 111, and a change factor variable display unit 112. It also has an adjustable variable processing unit 113 as a function to impose constraints on controlling the state of the manufacturing equipment 200.

[0043] The processing performed in the monitoring processing unit 114, which consists of a variable definition unit 103, a data formatting processing unit 104, and a principal component analysis processing unit 105, is a general principal component analysis, such as the one described in Patent Document 1.

[0044] The data storage unit 101 is a storage unit that stores (stores) the detection signal group output from the sensor group 300 as raw data for each time period. The data acquisition unit 102 acquires data from the sensor group 300 online for the time being monitored. The offline data storage unit 101 and the online data acquisition unit 102 can be considered functions of the data collection and management system 800 depending on the system configuration, but in this embodiment they are described as one of the functional blocks in Figure 4.

[0045] From the detection signal groups of the variable definition unit 103, data storage unit 101, and data acquisition unit 102, the detection signal group to be used for analysis is selected. For example, it is possible to select data from all sensors, or to select only the sensor group related to the reaction tank 203. In addition, the period to be used for analysis is defined from the time-series data of the data storage unit 101.

[0046] In this invention, the period of time-series data to be analyzed is shorter compared to conventional techniques for constructing learning models, etc. In this embodiment, although there is more than 200 days of data prior to the analysis period, the variable definition unit 103 uses a total of 20 days of data: 1 day of online data and 19 days of recent offline data. While a large amount of data is effective for constructing learning models, the conditions, such as climate and the state of consumables, may differ from those of the most recent analysis period. Therefore, in order to capture even slight changes in the state in the most recent period, the analysis period in this embodiment was fixed at 20 days, and the analysis data was updated each time an analysis was performed once a day.

[0047] The data formatting processing unit 104 performs preprocessing before inputting the data to the subsequent principal component analysis unit 105. For example, it unifies data sets acquired at different sampling times to a sampling time such as a 1-minute cycle or a 1-hour cycle. It also performs intensity normalization processing and removal of clearly abnormal data for each detection signal. After preprocessing the detection signal set, the principal component analysis processing unit 105 performs principal component analysis processing.

[0048] Principal component analysis transforms n sensor data points at each time point into a principal component space. The direction of the principal component axis 1, where the variance is greatest, is expressed by equation (1). k P is the sensor value after shaping. 1,k This is the coefficient for the k-th variable, which is used to convert it to the principal component axis 1 obtained by principal component analysis.

[0049]

number

[0050] Similarly, we calculate the second principal component score, Score2, which has the maximum variance in the direction orthogonal to the first principal component. By plotting Score1 on the x-axis and Score2 on the y-axis, we can plot the two-dimensional principal component scores. Expressing the score at each time t as an n-dimensional vector Z gives equation (2).

[0051]

number

[0052] The indicator definition unit 106 defines the productivity indicators to be analyzed. Examples include unit cost, yield rate, and energy efficiency. Alternatively, the analysis can be performed using the output of the predictive diagnostic system and anomaly detection system 600.

[0053] The clustering processing unit 107 classifies the indicators defined in the indicator definition unit 106 into multiple classes. While not a mandatory element, it is used when it is more appropriate than expressing productivity indicators as numerical values. For example, if the good product rate is the productivity indicator and the target is 80-90% for manufacturing equipment, classifying 90% or more as good, 80-90% as average, and below 80% as poor may be easier to understand than using numerical values. In such cases, clustering processing is applied to the productivity indicator.

[0054] Next, the state change display unit 108 will be described. As an example of display, a two-dimensional plot of principal component scores will be described. In the variable definition unit 103, 20 days' worth of data is acquired. There are n principal component scores, the same number as the variables, but it is appropriate to display them in two or three dimensions for display to the operator, and in this embodiment, it will be explained in two dimensions.

[0055] As described above, a typical method involves plotting a two-dimensional principal component score by plotting Score1 as PC1 on the horizontal axis and Score2 as PC2 on the vertical axis, and then determining normality and abnormality by focusing on these coordinates. The present invention, however, focuses not on each individual coordinate, but on the change in coordinates between two points, i.e., a vector, and displays this as information on the state change of the manufacturing equipment 200.

[0056] Figures 5A and 5B show typical display examples of the state change indicator unit 108 in this embodiment.

[0057] Figure 5A is an example of a two-dimensional plot, where the principal component analysis score is plotted as coordinate information, as in conventional methods. In this invention, each plot is further connected by a line in chronological order to indicate direction. Alternatively, information such as time and date can be added next to the plotted points. The output from the clustering processing unit 107 is also displayed overlaid. Specifically, in the case of the good product rate described above, if it is 90% or more and considered good, it is plotted with a circle; if it is between 80% and 90% and considered intermediate, it is plotted with a triangle; and if it is less than 80% and considered bad, it is plotted with an X.

[0058] This display may be based on the color of the plots, such as blue, yellow, or red, rather than the shape of the plots, or it may display ranks such as A, B, or C next to the plot points. Alternatively, the productivity indicator values ​​may be displayed directly without going through the clustering processing unit 107, and it is desirable to change this according to user convenience.

[0059] Display methods other than 2D plots will be explained using Figure 5B. The more variables included in a plot of the principal component axes, the more difficult it becomes to recognize the relationship with the actual equipment. This invention is characterized by performing principal component analysis processing while sequentially inputting online data. In this embodiment, the analysis is performed using data from the most recent 20 days, including the time of online data acquisition. Therefore, the principal component axes change each day as the analysis and display are updated. In other words, P in equation (1) 1,k or P in equation (2) nThis changes with each analysis and display. If the display is updated daily, the vertical and horizontal axes currently displayed will be different from those displayed the previous day.

[0060] This display with fluctuating axes can be difficult for users to understand and may feel cumbersome to use. Therefore, instead of the 2D plot in Figure 5A, the distance between two points in the 2D plot (Euclidean distance) and the direction can be plotted against the time axis.

[0061] The upper part of Figure 5B shows an example of plotting the distance (Euclidean distance) between two points over time. The middle part of Figure 5B plots the change in direction of two vectors defined between three points. Regarding the middle part of Figure 5B, a method is known in which principal component analysis is performed while sequentially updating the data. In this method, the amount of change in the principal component axis is focused on, and a threshold value is set for the amount of change, as in Patent Document 3, to achieve anomaly detection.

[0062] In comparison to the present invention, this method involves setting normal / abnormal thresholds and monitoring a time-series plot similar to the middle section of Figure 5B. On the other hand, the present invention visualizes changes in the state regardless of whether it is normal or abnormal, including the time-series plot in the middle section of Figure 5B, and visualizes indicators for extracting the factors causing the state change.

[0063] The distance in the upper part of Figure 5B is strongly related to the amount of change in the sensor group detection value, and the direction in the middle part of Figure 5B is strongly related to the type of variable that changed during the period. Therefore, if the direction of the plot changes and the distance is large, it is possible to inform the user (issue an alert) that the state of the manufacturing equipment 200 may have changed significantly.

[0064] For example, the transition distance from 5 / 10 to 5 / 11 is greater than 3, which is larger than the average value of 2 over the 20 days the data was collected. Also, the change in state from 5 / 10 to 5 / 11 is in a different direction than the change in state from 5 / 9 to 5 / 10. Here, as an example, 90°±20° is defined as a large change. Around 180°, although the sign of the variable is different, the type of variable extracted is similar to 0°, so it is a small angular change.

[0065] From the above conditions of the transition length and the angle change, a time series plot is shown in the lower part of FIG. 5B, where when the state change is large, it is set to 1, and when it is small, it is set to 0. For example, the change from May 19 to May 20 has a small transition distance. Although the angle change is 135° and is not small, in the method of the present invention, since a slight state change may be plotted as an angle change, the state change from May 19 to May 20 is regarded as relatively small.

[0066] When imposing the condition that both the transition distance and the angle change are large, when analyzing the state change in the past 20 days from May 20, May 7, 11, 17, and 19 are displayed as the days with large state changes. Symbols of clustered productivity may be overlaid on this display. The above is the display of the state change display unit 108 in this embodiment on May 20. A two-dimensional plot and a one-dimensional time series plot may be shown together, or only one of them may be shown.

[0067] Subsequently, the variable correlation display unit 109 will be described. It has a function of displaying the calculation result of the principal component analysis processing unit 105, and although it is not an essential element constituting the present invention, it may be displayed according to the application.

[0068] FIG. 6A is a two-dimensional plot of the same principal component scores as FIG. 5A. FIG. 6B is a factor plot of the variables used when plotting FIG. 6A. The n-dimensional matrix P shown in Equation (2) n For, it is equivalent to showing a vector (P 1,k, P 2,k ) composed of the elements of the first and second rows. When the number of variables is small, there is no problem with visibility even if it is three-dimensionally displayed as (P 1,k, P 2,k, P 3,k ).

[0069] Figure 6B allows us to understand the correlations between variables in the principal component space. For example, variables plotted at the same point exhibit similar behavior. Furthermore, the 5 / 20 data point plotted in the lower left of Figure 6A suggests that the factor in the lower left of Figure 6B is relatively large, while the factor in the upper right is large as a negative value. While Figure 6B is effective for understanding the state when there are few types of variables, it becomes difficult to identify important variables when dealing with tens or hundreds of variables, making it unsuitable as a method for supporting user operations.

[0070] Therefore, in this invention, by providing a state change definition unit 110 and a change factor variable calculation processing unit 111, variables that have been analyzed as important for the state change that the user has focused on from among many variables are output to the change factor variable display unit 112.

[0071] The function of the state change definition unit 110 in this embodiment is to select the day on which to display the factor variable. For example, one method is to automatically select 5 / 11, which is suggested to be a day with a large state change from the time series plot in Figure 5B. Alternatively, one method is to automatically select 5 / 20, which is the latest date. Alternatively, one method is to have the user check the time series plot screen in Figure 5A or Figure 5B and select the date on the screen.

[0072] Another method of operation in the state change definition unit 110 is shown in Figure 7. The user can visually confirm the state changes of the target of analysis in Figure 5A, which is displayed on the state change display unit 108 as a two-dimensional plot, and can also select them by clicking the pointer on the monitor screen. For example, if the user focuses on the state change from 5 / 10 to 5 / 11, they can select the line connecting 5 / 10 and 5 / 11 as shown by arrow (A), which can be expressed by equation (3). In this example, t = 5 / 11 and k = 1.

[0073]

number

[0074] Another method for defining state changes is to insert arrows into a two-dimensional plot. As shown in the variable correlation display unit 109, variables with high correlation can be displayed as factor vectors in the principal component space. Therefore, in order for the user to focus on the change from 5 / 10 to 5 / 11 and conduct analysis, the direction of the state change can be defined by defining an arrow (B) parallel to the transition direction from 5 / 10 to 5 / 11. For example, this method can be operated with a mouse on a monitoring device, or defined on a touch panel screen.

[0075] This method has the advantage of allowing the definition of any direction. Data selection based on dates and plot points only extracts factor variables for approximately 19 types of state changes in this example. On the other hand, by inserting arrows into the 2D plot, it is possible to investigate factor variables in directions that are not present in the data. Alternatively, one could select the start and end points of the arrows.

[0076] The operation of the state change definition unit 110 is useful not only for data analysis but also for preliminary studies to adjust plant parameters. For example, if production efficiency was relatively good on 5 / 20, it can be inferred that production efficiency will decrease if the state shifts to the right on the 2D plot. Therefore, the state change definition unit 110 can be used to identify variables with a large correlation in the left-right direction (C), i.e., the PC1 axis direction, in advance, and if there is a tendency for the state to shift to the right, preparations can be made to adjust the variables that move the state to the left. In addition, the inverse operation of the normalization process performed by the data shaping processing unit 104 can be performed from the length and direction of the arrows defined during principal component display, and an example of sensor detection values ​​based on actual physical quantities can be displayed.

[0077] The variable calculation processing unit 111 finds variables that have a high correlation with the state changes defined in the state change definition unit 110. In this embodiment, the n factor plots calculated by the principal component analysis processing unit 105 and, if necessary, output to the variable correlation display unit 109 are treated as vectors, and the dot product of the vectors defined in the state change definition unit 110 is calculated as shown in equation (4).

[0078]

number

[0079] The process also includes sorting the n dot product values ​​and displaying them in the variable change display unit 112. This sorting process may involve sorting by dot product value or absolute value of the dot product. Alternatively, there may be a process to select the 10 largest dot product values.

[0080] The variable change display unit 112 displays the results of the variable change calculation processing unit 111.

[0081] Figure 8 shows an example of displaying the dot product values ​​of all variables shown in equation (4) in order of process. Figure 9 shows an example of displaying the variables in order of the magnitude of their dot products. The white areas in the bar graph indicate that they behave the same as the variable to their left and can therefore be considered the same variable. Alternatively, you could display 10 to 20 variables with large absolute values ​​of their dot products as highly correlated variables, and variables with small absolute values ​​of their dot products as less correlated variables. In Figure 9, the data is displayed as a bar graph, but you could also display a table with the correlation values.

[0082] Furthermore, as an example of a method for extracting the factor variables, the result of multiplying the dot product value shown in equation (5) by the variation in each sensor value of the sensor group 300 may be displayed.

[0083]

number

[0084] As described above, the system supporting manufacturing in this embodiment includes a plurality of sensors (sensor group 300) that measure state quantities or manipulated quantities of the manufacturing equipment 200, a data acquisition and management unit (data acquisition and management system 800) that collects and stores time-series data of a plurality of variables measured by the plurality of sensors, a variable definition unit 103 that defines variables to be analyzed from the stored data and online measured variables, a data shaping processing unit 104 and a principal component analysis processing unit 105 that preprocess the variables defined in the variable definition unit 103 and perform principal component calculations, and a data shaping processing unit 104 that performs data calculations for a predetermined period at predetermined cycles according to the data. The system includes a variable correlation display unit 109 that performs multivariate analysis, updates the analysis values ​​based on the multivariate analysis for a predetermined period at predetermined cycles and displays them as time-series information, and displays the time-series information that serves as the target indicator in accordance with the time-series principal component values; a state change definition unit 110 that defines state changes that are the target of factor variable extraction from the time-series information based on the multivariate analysis; a change factor variable calculation processing unit 111 that performs a factor variable extraction process that has a high correlation with the state changes defined by the state change definition unit 110; and a change factor variable display unit 112 that displays the factor variables related to the state changes extracted by the factor variable extraction process.

[0085] Furthermore, the time-series information used as target indicators consists of actual productivity indicator data or predicted values ​​based on predictive diagnostics.

[0086] According to the manufacturing support system and manufacturing support method of this embodiment, even slight changes in the state of normally operating manufacturing equipment can be visualized. Furthermore, by extracting the factor variables of these state changes, it is possible to provide guidelines for adjusting equipment parameters to improve the production efficiency of the manufacturing equipment. In addition, it becomes possible to formalize parameter adjustments that previously relied on the experience of skilled technicians. [Examples]

[0087] Referring to Figures 10 to 13, a manufacturing support system and manufacturing support method according to Embodiment 2 of the present invention will be described.

[0088] Figure 10 shows an example of data from each part of the chemical plant management system in this embodiment. Figure 10 shows data for the period from March 1st to March 20th.

[0089] Productivity is displayed in the productivity data management system 400, and it has remained around 0.9 for the target period. Information from the predictive diagnostic system and anomaly detection system 600 based on principal component analysis can also be displayed. The monitoring and control unit 700 monitors the detection signals of each of the multiple sensors in the sensor group 300 and determines the abnormality of each individual sensor. If it is easier to understand the overall picture by displaying the productivity data management system 400, the predictive diagnostic system and anomaly detection system 600, the monitoring and control unit 700, and the plant monitoring and control support system 100 on the same display device, they may be displayed side by side as shown in Figure 10.

[0090] Although the predictive diagnostic system and anomaly detection system 600 are not essential components in the present invention, their details are described below for the purpose of comparing them with the configuration shown in Figure 4 as representing the present invention.

[0091] Figure 11 shows an example of a predictive diagnostic system and anomaly detection system 600 that determines whether or not there is an anomaly in the behavior of the detection signal group of sensor group 300 obtained through the data acquisition and management system 800.

[0092] The data storage unit 601 stores past detection values ​​from the sensor group 300 in order to build a learning model. Additionally, productivity data collected and analyzed by the productivity data management system 400 is also stored as offline data. Alternatively, the data storage unit 601 may utilize the functions of the data collection and management system 800.

[0093] In the variable definition unit 602, items necessary for constructing a learning model for process monitoring are selected from the sensor detection values ​​stored in the data storage unit 601, and new variables such as management indicators are synthesized as needed to define n input variables.

[0094] The data shaping processing unit 603 performs preprocessing on the input variables defined in the variable definition unit 602, including normalization and removal of outliers or data not used for training, before inputting the data to the subsequent principal component analysis processing unit 604.

[0095] The principal component analysis processing unit 604 calculates the direction in which the variance of the data is maximized from the centroid (first principal component), and then calculates the location in the direction orthogonal to the first principal component in which the variance is maximized (second principal component). This process is repeated for each data dimension. The above processing is the same as in the present invention, except that only offline data is used.

[0096] P in equation (2) n When learning is complete, P n(learning) It is fixed as and when processing online data, this P n(learning) Use this.

[0097] In the principal component score extraction processing unit 606, as an example, the relationship between the score of the first principal component analysis and the score of the second principal component analysis is created as a two-dimensional graph based on the results of the principal component analysis processing. This corresponds to the display of the predictive diagnostic system and anomaly detection system 600 in Figure 10. Information on productivity indicators defined in the indicator definition unit 605 may be added to the two-dimensional graph of scores.

[0098] If the plots of normal values ​​and abnormal values ​​can be separated for defined productivity, then an anomaly can be diagnosed. For example, in the display of the predictive diagnostic system and anomaly detection system 600 in Figure 10, if all "×: defective" can be plotted in the abnormal region, it means that the determination of normal / abnormal can be accurately made. In addition, the degree to which the observed data is far from the normal / abnormal boundary can be quantified as the degree of abnormality or normality. As a method of quantification, indicators such as the Q statistic and the T2 statistic can be used. These statistics are calculated by the statistic calculation unit 607, and the threshold definition unit 608 defines the threshold for abnormality, completing the model.

[0099] During monitoring, the data acquisition unit 609 selects the detected values ​​to be used for analysis from the online data acquired from the sensor group 300 via the data collection and management system 800, and the variable definition unit 610 selects the detected values ​​to be used for analysis, and preprocessing and P are performed using the conditions used during learning. n(learning) The principal component score is calculated from the data. Furthermore, after the data is formatted in the data formatting processing unit 611, statistics calculation processing unit 612 calculates statistics such as the degree of abnormality and the degree of normality. In addition, the abnormality determination processing unit 613 determines whether the statistics exceed the threshold defined in the threshold definition unit 608, and the diagnosis result of normal or abnormal is displayed in the status determination display unit 614.

[0100] In Figure 10, the predictive diagnostic system and anomaly detection system 600 have data for the analysis period (March 1st to March 20th) plotted in the area enclosed by the dotted line. Based on learning from past data, anomalies are defined as being in the lower right area, and there are no signs of moving into the anomaly area in the subsequent period. The monitor and control unit 700 screen displays the sensor values. In reality, approximately 100 sensor detection values ​​are monitored, but here only one detection value each from the flow sensor, pressure sensor, concentration sensor, and temperature sensor attached to the reaction tank 203 is displayed.

[0101] Figure 10 shows the information from the state change display unit 108 on the screen of the plant monitoring and control support system 100. Principal component analysis based on this 20-day data revealed a significant state change from March 14th to March 15th. Furthermore, the 2D plot allows for the identification of the state on March 19th and March 20th, which were periods of good productivity, suggesting that the state change from March 14th to March 15th had a positive impact. By extracting and understanding the factors behind the state change from March 14th to March 15th, users can obtain guidance for parameter adjustments from March 20th onward.

[0102] By selecting 3 / 15 on the screen of the plant monitoring and control support system 100 shown in Figure 10, the state change to be analyzed is defined.

[0103] Figure 12 shows the changes in sensor values. The upper part of Figure 12 shows the results of extracting the sensors with large normalized changes in the detected values ​​from the sensor group 300 installed in the raw material tank 201 and reaction tank 203 from March 14 to March 15. Furthermore, the lower part of Figure 12 shows the results of multiplying by the dot product of the factor variables extracted from principal component analysis using equation (5), indicating that although the changes in the detected values ​​of A_Upper Pressure 3 and A_Upper Pressure 4 were large in this state change, their contribution to the state change was small.

[0104] The table on the right in Figure 12 shows the top 10 variables that influenced the state change from March 14th to March 15th, extracted from the sensor value change and the dot product of the principal component analysis. From this table, the user can interpret the state change from March 14th to March 15th and obtain information that will serve as a guide for future productivity improvements. Note that although the table on the right in Figure 12 shows normalized values, it is also possible to display an example of sensor detection values ​​based on actual physical quantities by performing the inverse operation of the normalization process performed by the data formatting processing unit 104.

[0105] Figure 13 conceptually illustrates the differences in user behavior expected depending on whether or not the present invention is applied.

[0106] The change in state from 3 / 14 to 3 / 15 was an expected outcome for skilled workers, and adjustments to the manufacturing equipment 200 were possible based on past knowledge. However, the conventional operation using the monitoring and control unit 700 shown in the left diagram of Figure 13 may be insufficiently understood by users with little operational experience.

[0107] On the other hand, as shown in the right-hand figure of Figure 13, the application of the plant monitoring and control support system 100 of the present invention can contribute to improved understanding and acceptance by using data when explaining to less experienced users, for example, for the purpose of passing on experience. Alternatively, the operation of the manufacturing equipment 200 can be carried out while sharing a hypothesis among multiple users that the change in state from 3 / 14 to 3 / 15 had a positive effect.

[0108] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are described in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace parts of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add configurations from other embodiments to the configuration of one embodiment. In addition, it is possible to add, delete, or replace parts of the configuration of each embodiment with other configurations.

[0109] Furthermore, each of the above configurations, functions, processing units, and processing means may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. Alternatively, each of the above configurations and functions may be implemented in software by having the processor interpret and execute programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. [Explanation of Symbols]

[0110] 1…Chemical plant management system 100... Plant monitoring and control support system 101,601...Data storage unit (offline) 102,609...Data acquisition unit (online) 103,602,610… Variable definition section 104,603,611…Data formatting processing unit 105,604... Principal component analysis processing unit 106…Indicator definition part 107...Clustering Processing Unit 108...Status change display unit 109...Variable correlation display section 110...State change definition section 111... Processing unit for calculating variable factors of change 112...Display section for variable factors causing change 113... Adjustable Variable Processing Unit 114... Monitoring Processing Unit 200…Manufacturing equipment 201... Raw material tank 202...Storage tank 203, 204… reaction tanks 205…Product storage tank 300... Sensor group 400… Productivity Data Management System 500…Manufacturing Management System 600…Predictive diagnostic system and anomaly detection system 605...Indicator definition section 606... Principal component score extraction processing unit 607…Statistics calculation unit 608...Threshold definition section 612...Statistical calculation processing unit 613... Anomaly detection processing unit 614...Status determination display unit 700... Monitoring and Control Unit 800...Data collection and management system (data collection system)

Claims

1. Multiple sensors for measuring the state or operation of manufacturing equipment, A data acquisition and management unit that collects and stores time-series data of multiple variables measured by the multiple sensors, A variable definition unit defines the variables to be analyzed from the retained data and online measured variables, A data formatting processing unit and a principal component analysis processing unit that preprocess the variables defined in the variable definition unit and perform principal component calculations, A variable correlation display unit that performs multivariate analysis according to data for a predetermined period at predetermined cycles, updates the analysis values ​​based on the multivariate analysis for the predetermined period at predetermined cycles and displays them as time-series information, and displays the time-series information that serves as the target indicator in accordance with the time-series principal component values, A state change definition unit defines state changes that are the target of factor variable extraction from the time series information based on the multivariate analysis, A change factor variable calculation processing unit that performs a process to extract factor variables that have a high correlation with the state change defined by the state change definition unit, A change factor variable display unit that displays the factor variables related to the state change extracted by the factor variable extraction process, A manufacturing support system characterized by comprising the following features.

2. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized by updating analytical values ​​based on multivariate analysis over a predetermined period at predetermined cycles and displaying them as time-series information, as well as displaying the relationships between variables as a factor plot.

3. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized by displaying numerical data of time-series information that serves as the target indicator as a class classified by clustering, in accordance with the time-series principal component values.

4. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized in that, when defining state changes to be targeted for factor variable extraction from the time series information based on the multivariate analysis, two arbitrary points are selected from the analysis values ​​based on the multivariate analysis for a predetermined period for each predetermined cycle.

5. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized in that, when defining state changes that are the target of factor variable extraction from the time series information based on the multivariate analysis, the analysis values ​​based on the multivariate analysis for a predetermined period are updated at predetermined cycles, and the direction of the state change is input on the screen displayed as time series information, thereby defining the state change.

6. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized by performing a factor variable extraction process that is highly correlated with state changes, by calculating the dot product of the factor vector representing the variable relationship in the principal component space and the state change vector that is the target of factor variable extraction defined from the time series information, for each variable.

7. A system for supporting the manufacturing described in claim 6, A manufacturing support system characterized by multiplying the dot product of a factor vector representing the variable relationship in the principal component space and the state change vector that is the target of factor variable extraction defined from the time series information by the amount of change of the preprocessed variable, in a process for extracting factor variables that are highly correlated with state changes.

8. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized in that, for time-series information in which analytical values ​​based on multivariate analysis for a predetermined period are updated at each predetermined cycle, the analytical values ​​based on multivariate analysis are displayed as the Euclidean distance between the plot of principal component analysis at the current time and the time one cycle prior.

9. A system for supporting the manufacturing described in claim 8, A manufacturing support system characterized by issuing an alert when the Euclidean distance is greater than a predetermined value.

10. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized by displaying factor variables that have a high correlation with the extracted state changes.

11. A system for supporting the manufacturing described in claim 1, A manufacturing support system characterized in that the time-series information used as the target indicator is either actual productivity indicator data or predicted values ​​based on predictive diagnostics.

12. A system for supporting the manufacturing described in claim 2, A manufacturing support system characterized by converting the magnitude of the correlation of the extracted state change factors, based on principal component analysis, and the defined state change amount into physical quantities of sensor-detected values ​​and displaying them.

13. A method for supporting manufacturing using a system for supporting manufacturing, (a) A step in which multiple sensors measure the state or manipulated quantity of the manufacturing equipment, (b) The data collection and management unit collects and stores time-series data of multiple variables measured by the multiple sensors, (c) The variable definition unit defines the variables to be analyzed from the retained data and the online measured variables, (d) The data formatting processing unit and the principal component analysis processing unit preprocess the variables defined in the variable definition unit and perform principal component calculations, (e) The variable correlation display unit performs multivariate analysis according to the data for a predetermined period at predetermined cycles, updates the analysis values ​​based on the multivariate analysis for the predetermined period at predetermined cycles and displays them as time series information, and displays the time series information that serves as the target indicator in accordance with the time series principal component values. (f) The state change definition unit defines state changes that are the subject of factor variable extraction from the time series information based on the multivariate analysis, (g) The change factor variable calculation processing unit performs a process to extract factor variables that have a high correlation with the state change defined in the state change definition unit, (h) The change factor variable display unit displays the factor variables related to the state change extracted by the factor variable extraction process, A method for supporting manufacturing, characterized by including the following.

Images