Control device and control method

Abstract

To enable the presentation of a basis of control operation to an object subjected to control.SOLUTION: A control device 100 includes: a state calculating unit 112 that calculates the state of an object subjected to control on the basis of the state quantity of the object subjected to control; an operation amount calculating unit 114 that calculates a control operation amount for controlling the object subjected to control on the basis of the state; and a display control unit 115 that displays the state of the object subjected to control and the control operation amount on a display apparatus. The state of the object subjected to control displayed by the display control unit 115 includes a time-course change in the state of the object subjected to control after the control by the control operation amount is executed.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a control device and a control method for controlling equipment and devices.

Background Art

[0002] In recent years, plant control systems have implemented control logic not only for normal continuous operation but also for unsteady operations such as startup, shutdown, raw material switching, and when an abnormality occurs, and the automation of operation has advanced. However, for the control of unsteady operations, stable operation has not necessarily been achieved, and abnormalities such as the temperature exceeding the alarm level may occur during operation. Therefore, in unsteady operations, the operator monitors the main sensor values of the plant, and when it is determined that the sensor values deviate from the allowable range, an intervention operation on the control logic is executed.

[0003] This intervention operation refers to an operation of changing an operation amount such as the flow rate controlled by the control system through the operation monitoring panel. For example, if it is determined that the temperature is likely to reach the alarm level as a result of the temperature rising during unsteady operation, the operator performs an operation to increase the cooling flow rate in order to lower the temperature.

[0004] Even in such a case, the cooling flow rate is adjusted to a predetermined value by the control system, but the operator makes a slight change to this reference value. There are buttons on the operation monitoring panel for adding or subtracting from the reference value, and each time a button is pressed, an intervention operation can be performed such that the flow rate increases, for example, by 1.0 t / h with respect to the reference value. Thus, the value for changing with respect to the reference value output by the control logic is called a bias value.

[0005] As explained above, during non-steady-state operation, if the sensor value is determined to be outside the normal range, the operator intervenes in the operation by manipulating the bias value to bring the sensor value back within the normal range. It is also possible to manipulate the bias value during normal continuous operation, but since the sensor value is usually within the normal range during normal operation, such intervention is unnecessary.

[0006] However, the stability of operation during non-routine operations depends on the operator's skill level. Skilled operators can ensure a certain level of stability during non-routine operations, including starting, stopping, material switching, and handling abnormal situations. However, non-routine operations performed by unskilled operators often result in frequent alarm activations and long startup times, making them less reliable and economical than those performed by skilled operators. Although there are regular forums for discussing operational know-how between skilled and unskilled operators, the reality is that the transfer of operational know-how is not progressing because even skilled operators often operate based on tacit knowledge.

[0007] In light of these on-site challenges, attempts are being made to prevent the personalization of operations by using artificial intelligence technology to learn the tacit knowledge of skilled operators and constructing control models based on that knowledge. However, one problem when applying artificial intelligence technology to plant control is the lack of explainability regarding the output of the control model. Even if a control model can be constructed using artificial intelligence technology, if the control model is a black box and operators cannot judge the appropriateness of the control quantities output by the control model, the introduction of artificial intelligence technology will be difficult. Thus, in control systems that utilize artificial intelligence technology, it is desirable to also present operators with the basis for the judgment of the control quantities output by the control model.

[0008] To address the above challenges, an example of a control system that provides a basis for decision-making is the water treatment plant control system described in Patent Document 1. The water treatment plant control system is a sewage treatment plant that uses clustering and fuzzy logic to identify the operating state and includes means for selecting control setting values ​​according to the operating state. The embodiment describes a process that identifies weather conditions using sensor values ​​from the plant, identifying states such as sunny, initial rainfall, and prolonged rain, and selecting control setting values ​​according to the state. In parallel with the control of the plant, the operator can confirm the operating state identified by the control device through an operating state notification means. In other words, when the control system selects control setting values, it presents the weather conditions that serve as the basis for that selection. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] Japanese Patent Publication No. 2003-140712 [Overview of the project] [Problems that the invention aims to solve]

[0010] During non-steady-state operations such as startup, shutdown, raw material switching, and abnormal occurrences, sensor values ​​(hereinafter also referred to as state variables) may temporarily deviate from the normal range, triggering alarms. However, if the impact is minor, feedback control by the normal control system will eventually return the sensor values ​​to the normal range, and the plant will converge to a stable state. Nevertheless, deviations from the normal range of sensor values ​​and the triggering of alarms reduce operational safety, so operators perform non-steady-state operations to avoid this. In this case, the plant's components transition through several states before finally reaching a stable state. An example is shown below.

[0011] If the temperature rises beyond the normal range, the cooling flow rate increases by opening the valve that adjusts the cooling flow rate. If the pressure also rises at this time, the discharge flow rate increases by opening the valve that adjusts the discharge flow rate to reduce the pressure. Next, the temperature begins to decrease due to the increase in cooling flow rate. Similarly, the pressure begins to decrease due to the increase in discharge flow rate. Once the temperature and pressure approach the set values ​​to a certain extent, the valve openings that adjust them also converge to constant values. Through these stages, the temperature, flow rate, and pressure finally stabilize. In this final state, some state variables temporarily increase and then return to their original values ​​over time, while other state variables transition to other state variables that balance the temperature rise. As described above, several operating states are reached before the final stabilization, and the appropriate control variables differ for each state.

[0012] As explained above, in transient operation, after an abnormal state occurs, the system undergoes several state transitions and corresponding control operations repeatedly before finally transitioning to a stable state. Skilled operators have mastered this entire process of state transitions and are capable of performing control operations according to the state.

[0013] The water treatment plant control system described in Patent Document 1 did not identify and present each operating state to the operator for each transition state the plant takes until it finally reaches a stable state, nor did it present the rationale for determining the appropriate control operation amounts for each state. Even if the control system presented control operations for each state, an unskilled operator who may not have mastered the entire state transition would perform the control operations without understanding the rationale (reason) for doing so.

[0014] This invention was made in view of the above background, and aims to provide a control device and control method that enable the presentation of the basis for control operations. [Means for solving the problem]

[0015] To solve the above problems, the control device according to the present invention includes a state calculation unit that calculates the state of the control target based on the state quantity of the control target, an operation quantity calculation unit that calculates a control operation quantity for controlling the control target based on the state, and a display control unit that displays the state of the control target and the control operation quantity on a display. The state of the control target displayed by the display control unit includes the change over time of the state of the control target after the control of the control operation quantity is performed. The system further includes an abnormality detection unit that detects an abnormality in the controlled object by referring to a normal state database that stores the normal range of the state quantities of the controlled object. The manipulated amount calculation unit calculates the control manipulated amount when an abnormality in the controlled object is detected. The display control unit refers to a state transition database that stores state transitions showing the change in state over time from an initial state, which is the state of the controlled object in which the abnormality was detected, to a steady state, which is a state in which the change in the state quantities of the controlled object is less than or equal to a predetermined value, and displays the change in state over time included in the state transition, with the abnormality detected by the abnormality detection unit as the initial state.

Effect of the Invention

[0016] According to the present invention, it is possible to provide a control device and a control method that enable the presentation of the basis of control operations. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

Brief Description of the Drawings

[0017] [Figure 1] It is an overall configuration diagram of the plant control system according to the present embodiment. [Figure 2] It is a functional block diagram of the operation monitoring control system according to the present embodiment. [Figure 3] It is a functional block diagram of the control device according to the present embodiment. [Figure 4] It is a data configuration diagram of the plant state quantity database according to the present embodiment. [Figure 5] It is a data configuration diagram of the state database according to the present embodiment. [Figure 6] It is a data configuration diagram of the state transition database according to the present embodiment. [Figure 7] It is a flowchart of the control process according to the present embodiment. [Figure 8] It is a screen configuration diagram of the state progress display screen according to the present embodiment. [Figure 9] It is a screen configuration diagram of the trend graph display screen according to a modification of the present embodiment. [Figure 10] It is a screen configuration diagram of the trend graph display screen according to a modification of the present embodiment. [Figure 11]It is a screen configuration diagram of a trend display screen shown in a system diagram according to a modification example of the present embodiment. [Figure 12] It is a screen configuration diagram of a state progress display screen according to a modification example of the present embodiment.

Mode for Carrying Out the Invention

[0018] ≪Outline of Control Device≫ The control device in the mode (embodiment) for carrying out the present invention will be described below. The control device acquires measurement values (sensor signals, state quantities) of sensors installed in facilities or devices constituting a plant (control target) via an operation monitoring control system. The control device performs clustering processing on the state quantities to identify the state of the plant. In other words, the control device regards those with similar state quantities as vector values as one state, and identifies the state of the plant from the state quantities.

[0019] The control device includes a state transition database (see FIG. 6 described later) that stores a state transition, which is a time-series change of the state from when the state quantity deviates from the normal state until the plant returns to the steady state. Further, the control device calculates (estimates) a control operation quantity from the current state and the past state using a control operation quantity calculation model, calculates the control operation quantity after deviating from the normal state, and transmits it to the operation monitoring control system. By the operation monitoring control system executing the control of the control operation quantity or the operator performing the operation of the control operation quantity, the plant returns to the steady state. The control operation quantity calculation model is a machine learning model that has learned the control operation quantity performed by a skilled operator. The control operation quantity is the bias value described later.

[0020] The control unit transmits not only the manipulated variable but also the state transitions until the plant reaches a steady state to the operation monitoring and control system. The operation monitoring and control system displays both the state transitions and the manipulated variable. Operators can then see how the plant progresses to a steady state through their control operations on the manipulated variable, providing a basis for their control operations. Consequently, operators can perform manual control operations more smoothly. Furthermore, operators can learn the control operations performed by experienced operators.

[0021] ≪Overview of the controlled system (plant)≫ Figure 1 is an overall configuration diagram of the plant control system 10 according to this embodiment. The plant control system 10 consists of a plant 300, an operation monitoring and control system 200, and a control device 100. The plant 300 is the controlled object monitored and controlled by the operation monitoring and control system 200 and the control device 100. In the plant 300, raw material 311 is fed into the column 315 and reacts to obtain a product. A raw material supply flow rate control valve 313 is installed in the raw material supply line, and the flow rate is controlled to a predetermined value (set value) by operating the valve opening. In addition, a pressure sensor 312 for measuring the raw material supply pressure and a flow sensor 314 for measuring the raw material supply flow rate are installed as sensors. A pressure sensor 316 for measuring the internal pressure of the column 315 is also installed.

[0022] In this embodiment, the processing of the plant control system 10 will be explained using an example of an abnormal operation. The abnormal operation performed by the plant 300 operator aims to quickly restore the internal pressure (in-column pressure) of the column 315, which has risen abnormally due to an abnormal increase in the supply pressure of the raw material 311, back to its original state. In this example, even without performing an abnormal operation, feedback control to the valve opening of the control valve 313 works so that the in-column pressure follows the set value (target value), so the in-column pressure eventually returns to its original state (set value of in-column pressure, control target value). However, by performing an abnormal operation that temporarily lowers the set value of the in-column pressure, the response operation of the valve opening is accelerated, and the time it takes for the in-column pressure to return to its original state can be shortened. If the in-column pressure remains abnormally high, there is a possibility of fatal damage to the equipment, so it is desirable to lower the pressure as quickly as possible, and an abnormal operation is desirable.

[0023] When an operator performs such an unsteady operation manually, they change the bias value relative to the set value of the tower pressure via the screen of the monitoring operation panel 203 (see Figure 2 below) provided in the operation monitoring and control system 200. The screen has buttons, for example, to change the bias value. When a button is pressed once, the bias value is changed by a specified amount from the current value. In this example, the set value is to be temporarily reduced, so the operator presses the button that decreases the bias value a number of times corresponding to the amount of reduction. If the bias value before the operation is 0.0, pressing the button makes the bias value negative. Because a negative bias value is added, the set value decreases compared to before the operation. Next, while observing the situation in which the tower pressure is recovering to its original state, the operator presses the button that increases the bias value the same number of times to return the bias value to the original 0.0. As a result, the set value returns to its original value.

[0024] The control device 100 according to this embodiment replaces the non-steady-state operations that are manually performed by the operator as described above. The control device 100 calculates an appropriate control operation amount (bias value) according to the state of the plant 300 and transmits it to the operation monitoring and control system 200. The operation monitoring and control system 200 adds the bias value to the set value (in the example above, the pressure inside the tower) and performs feedback control.

[0025] <<Overview of the Operation Monitoring and Control System>> Figure 2 is a functional block diagram of the operation monitoring and control system 200 according to this embodiment. The operation monitoring and control system 200 comprises a plant state variable transmission / reception unit 201, a plant control unit 202, and a monitoring operation panel 203. The plant state information transmission / reception unit 201 receives measured values ​​(state information) from pressure sensors 312, 316, flow sensor 314, and other sensors in the plant, and transmits them to the control device 100. The plant state information transmission / reception unit 201 also transmits the set values ​​of the state information. The state information and set values ​​are transmitted at predetermined intervals. In this embodiment, they are transmitted at 10-minute intervals.

[0026] The plant control unit 202 monitors state variables based on the implemented control logic and outputs signals to control (feedback control) the equipment and devices in the plant 300. For example, the plant control unit 202 outputs a signal to control the opening degree of the control valve 313 so that the internal pressure of the tower 315 becomes a predetermined pressure (a set value for the state variable called internal pressure). When the plant control unit 202 receives a control operation variable (bias value) from the control device 100, it adds the bias value to the initial set value to create a new set value and performs feedback control based on this set value.

[0027] The monitoring and operation panel 203 displays not only state variables but also transition states and control operations received from the control device 100 (see Figure 8 below). The operator can set a bias value for the set value of a state variable by operating the monitoring and operation panel 203. For example, if a negative bias value is set for the set value of the internal pressure of the tower 315, the plant control unit 202 controls the opening of the regulating valve 313 to lower the internal pressure. Since setting the bias value controls the plant 300 via the plant control unit 202, the bias value is also referred to as the control operation variable.

[0028] ≪Overall configuration of the control system≫ Figure 3 is a functional block diagram of the control device 100 according to this embodiment. The control device 100 is a computer and comprises a control unit 110, a storage unit 120, and an input / output unit 180. User interface devices such as a display, keyboard, and mouse are connected to the input / output unit 180. The input / output unit 180 is equipped with a communication device and is capable of sending and receiving data with the operation monitoring and control system 200.

[0029] The control unit 110 includes a CPU (Central Processing Unit) and comprises a state quantity receiving unit 111, a state calculation unit 112, an anomaly detection unit 113, an operation quantity calculation unit 114, and a display control unit 115. The memory unit 120 is comprised of memory devices such as ROM (Read Only Memory), RAM (Random Access Memory), and SSD (Solid State Drive). The memory unit 120 stores a plant state variable database 130 (see Figure 4 below), a state database 140 (see Figure 5 below), a state transition database 150 (see Figure 6 below), a normal state database 160, a control operation variable calculation model 121, and a program 128. Program 128 includes a description of the control process shown in Figure 7 below. The configuration of the control unit 110 and the storage unit 120 will be described below.

[0030] ≪Status Indicator Receiving Unit and Plant Status Indicator Database≫ The state quantity receiving unit 111 stores the state quantities (measured values ​​from sensors installed in the plant 300) and the set values ​​of the state quantities transmitted by the operation monitoring and control system 200 in the plant state quantity database 130. Figure 4 is a data configuration diagram of the plant state quantity database 130 according to this embodiment. The plant state quantity database 130 is, for example, tabular data, where one row (record) indicates the state quantities and set values ​​received at the same time (within a predetermined time). In addition to the date and time of receipt, the record includes columns (attributes) for raw material supply pressure, tower pressure, raw material supply flow rate, raw material flow rate control valve opening, and tower pressure set value. The plant state quantity database 130 may include even more attributes of state quantities and the set values ​​of the state quantities.

[0031] Each record in the plant state database 130 indicates the state of plant 300 at the time indicated by the date and time. For example, the first record in the plant state database 130 shown in Figure 4 represents the state of plant 300 at 10:10 on September 3, 2021, as a five-dimensional vector (hereinafter also referred to as a state vector) with raw material supply pressure of 1.1 MPa, tower pressure of 0.8 MPa, raw material supply flow rate of 1.2 t / h, raw material flow control valve opening of 80%, and tower pressure setpoint of 0.8 MPa.

[0032] ≪Status Calculation Unit and Status Database≫ The state calculation unit 112 calculates the cluster number of the state quantity vector shown in each record of the plant state quantity database 130. Specifically, clustering is performed on the state quantity vectors accumulated in the past to classify them into clusters. In other words, state quantity vectors that are close in distance are classified to belong to the same cluster. Note that the state quantities may be normalized during the clustering process. Each cluster is assigned a cluster number as identification information. The state calculation unit 112 calculates the cluster closest to the state quantity vector of the state quantity received by the state quantity receiving unit 111 and stores the cluster number of that cluster in the state database 140 (see Figure 5 below).

[0033] Furthermore, the clustering process, which classifies previously accumulated state vectors into clusters, does not need to be performed by the control device 100. The clustering process may be performed by a device that can access the accumulated state quantities of the plant 300, similar to the plant state quantity database 130.

[0034] Figure 5 is a data structure diagram of the state database 140 according to this embodiment. The state database 140 is, for example, tabular data, where each row (record) includes the attributes of date and time 141 and cluster number 142, indicating the cluster number 142 of the cluster to which the state vector at date and time 141 belongs. Cluster number 142 is identification information for clusters of similar state vectors, and it can be said that the state of plant 300 can be identified by cluster number 142. In other words, the state of plant 300 is represented by a state vector, and state vectors that are close in distance (clusters of state vectors, similar states) are identified by cluster number 142. Note that multiple clusters that are close in distance may be considered as a single state. The state database 140 can also be seen as showing the time-series changes in the state of plant 300 identified by cluster number 142.

[0035] As explained above, the state calculation unit 112 of the control device 100 calculates the state of the controlled object (plant 300, or equipment and devices installed in plant 300) based on the state quantities (sensor values) of the controlled object.

[0036] Anomaly detection unit The abnormality detection unit 113 detects abnormalities in the plant 300 by referring to the normal state database 160 based on the latest state quantities stored in the plant state quantity database 130. For example, the normal state database 160 stores the normal range of state quantities, and the abnormality detection unit 113 determines that there is an abnormality if any of the latest state quantities are outside the normal range. Alternatively, for example, the normal state database 160 stores cluster numbers that are determined to be abnormal, and the abnormality detection unit 113 may determine that there is an abnormality if the latest cluster number calculated by the state calculation unit 112 is a cluster number that is determined to be abnormal.

[0037] As described above, the abnormality detection unit 113 in the control device 100 detects abnormalities in the controlled object (plant 300) by referring to the normal state database 160, which stores the normal range of the state variables of the controlled object.

[0038] ≪State transition database and state transition database≫ Figure 6 is a data structure diagram of the state transition database 150 according to this embodiment. The state transition database 150 includes a state transition table 151 that shows the state transitions, which are the time-series changes in the state of the plant 300. One state transition table 151 shows the state transitions starting from after an anomaly is detected in the plant 300 until it reaches a steady state (stable state). Each row (record) of the state transition table 151 indicates a state and includes the cluster number 152 of the cluster included in that state and the state name 153, and the state changes (transitions) sequentially from the top record to the bottom record.

[0039] The state transition table 151 shown in Figure 6 represents the state transitions from the occurrence of an abnormal rise in raw material supply pressure to the point where the pressure inside the tower returns to normal and stabilizes. Stabilization means that the change in each state variable of the plant 300 falls below a predetermined value, and the state becomes steady. Also, if the change in one state variable falls below a predetermined value and stabilizes, that state variable is also described as stabilized. The states (records) in the state transition table 151 are described in order below. Records with cluster number 152, "1-3," indicate an abnormal increase in raw material supply flow rate due to an abnormal increase in raw material supply pressure. Records with cluster number 152, "4-6," indicate that the pressure inside column 315 is rising due to an increase in the raw material flowing into column 315.

[0040] When the pressure inside the tower rises, the plant control unit 202 (see Figure 2) performs feedback control to ensure that the pressure inside the tower follows the set value. To control the pressure inside the tower, the plant control unit 202 adjusts the opening of the raw material supply flow control valve (see control valve 313 shown in Figure 1). Here, in order to lower the pressure inside the tower, the plant control unit 202 operates the raw material flow control valve opening in the downward direction (closing direction) to reduce the raw material supply flow rate. This state is represented by records with cluster number 152 set to "7-20". Records with cluster number 152, specifically "21-28," indicate that the raw material supply flow rate began to recover to its original state as the raw material flow control valve opening decreased. Records with cluster number 152, specifically "29-35," indicate that the column pressure also began to recover to its original state as the raw material supply flow rate recovered to its original state.

[0041] After a while, as indicated by the records for cluster number 152, "36-40", the raw material flow control valve opening will settle at a point where it is balanced. At this time, the raw material flow control valve opening will not return to its original state, but will settle at an opening that restores the raw material supply flow rate to its original value in response to the abnormal increase in raw material supply pressure, that is, at an opening lower than before the abnormality occurred. Next, as indicated by records "41-42" for cluster number 152, the raw material flow rate is stabilized by the adjustment of the raw material flow control valve opening, restoring the system to its original state. Finally, as indicated by records "43-45" for cluster number 152, the pressure inside the tower is restored to its original state.

[0042] The state transition table 151 is generated based on the plant state quantity database 130 that has been accumulated in the past. Specifically, the cluster number of the state quantity vector from the time an anomaly is detected until it is settled is calculated. Among the multiple clusters arranged in chronological order, cluster groups that are close in distance (have small differences in state quantities) are treated as a single state and become a single record in the state transition table 151. The cluster number 152 of the record is the cluster number of the cluster included in the cluster group. The state name 153 is set based on the name of the state quantity with the largest ratio of change in value among the state quantities of the cluster group and the direction of change.

[0043] Furthermore, the process of generating the state transition table 151 does not need to be performed by the control device 100. The process of generating the state transition table 151 may be performed by a device that can access the stored state quantities of the plant 300, similar to the plant state quantity database 130. Also, the state name 153 may be set manually rather than by the device.

[0044] <<Operational Variable Calculation Unit and Display Control Unit>> When the anomaly detection unit 113 detects an anomaly, the manipulated variable calculation unit 114 calculates the manipulated variable (bias value) based on clusters representing the state variable vectors of the plant 300 in the past within a predetermined time period. More specifically, the manipulated variable calculation unit 114 calculates the manipulated variable using a manipulated variable calculation model 121 that has learned the manipulated variable of a skilled operator. The manipulated variable calculation model 121 is a machine learning model in which the explanatory variable is the cluster number representing the state variable vector of the plant 300 in the past within a predetermined time period, and the dependent variable is the manipulated variable. The past within a predetermined time period refers to, for example, the present time, 10 minutes ago, 20 minutes ago, or 30 minutes ago.

[0045] The training data for the control variable calculation model 121 is constructed from the control operation history of skilled operators. The explanatory variables of the training data are the cluster numbers of the state variable vectors in the past within a predetermined time period at each point in a 10-minute cycle from the time an abnormality is detected in plant 300 until it settles down, and the dependent variable is the control variable (bias value) set by the skilled operator at each point in time. By using the control operation variable calculation model 121 generated by learning from such training data, the operation variable calculation unit 114 can calculate control operation variables similar to those of a skilled operator.

[0046] The type of control variable calculation model 121 is not limited; for example, it may be a neural network model, a support vector regression model, or a multiple regression analysis model. Furthermore, the process of preparing training data and generating the control variable calculation model 121 does not need to be performed by the control device 100. The generation process of the control variable calculation model 121 may be performed by a device that can access the stored state variables of the plant 300, similar to the plant state variable database 130. For example, a device that performs clustering of state variable vectors or generation of the state transition database 150 may also perform the generation process of the control variable calculation model 121.

[0047] As described above, the control variable calculation unit 114 in the control device 100 calculates a control variable for controlling the controlled object based on the state (cluster number indicating the state of the plant 300). The control variable calculation unit 114 also calculates a control variable when an abnormality is detected in the controlled object by the abnormality detection unit 113. Furthermore, the control variable calculation unit 114 calculates the control variable based on the state at the time of calculation and the state at a time prior to that calculation.

[0048] When the display control unit 115 detects an abnormality in the plant 300, it transmits the state name 153 from the state transition table 151 corresponding to the abnormality in the state transition database 150, and the control operation amount calculated by the operation amount calculation unit 114, to the operation monitoring control system 200. The monitoring operation panel 203 of the operation monitoring control system 200 displays the state name and the control operation amount (see Figure 8 below). The state transition table 151 corresponding to the abnormality is the state transition table 151 in which the cluster number 152 of the first record includes the cluster number 152 at the time the abnormality was detected (see cluster number 142 in Figure 5).

[0049] As described above, the display control unit 115 of the control device 100 displays the state of the controlled object and the controlled variable on the display unit (monitoring operation panel 203). The display control unit 115 also refers to the state transition database 150, which stores state transitions (state transition table 151) that show the changes in state over time from the initial state, which is the state of the controlled object when an abnormality is detected, to the steady state, which is the state where the change in the state variable of the controlled object is less than or equal to a predetermined value, and displays the changes in state over time that are included in the state transitions, with the abnormality detected by the abnormality detection unit 113 as the initial state (the first record in the state transition table 151).

[0050] Control Processing Figure 7 is a flowchart of the control process according to this embodiment. The control process includes the state acquisition process in steps S11 to S12 and the state progress display process in steps S21 to S28, which are executed in parallel.

[0051] In step S11, the status quantity receiving unit 111 receives the status quantity and set value transmitted by the operation monitoring and control system 200 and stores them in the plant status quantity database 130. In step S12, the state calculation unit 112 calculates the cluster number of the cluster that is closest to the state quantity vector, which is composed of the state quantity and set value received in step S11, and stores it in the state database 140.

[0052] In step S21, the anomaly detection unit 113 obtains the latest state values ​​from the plant state value database 130 and determines whether or not there is an anomaly. If there is an anomaly (step S21 → YES), the anomaly detection unit 113 proceeds to step S22; if there is no anomaly (step S21 → NO), it returns to step S21.

[0053] In step S22, the display control unit 115 identifies a state transition table 151 in the state transition database 150 that corresponds to the state quantity vector at the time when an abnormality was determined in step S21. Specifically, the display control unit 115 obtains the cluster number 142 at the time when an abnormality was determined from the state database 140, and identifies a state transition table 151 that includes this cluster number as the cluster number 152 of the first record. This state transition table 151 shows the state transitions from the time an abnormality was determined in step S21 until the plant 300 settles down.

[0054] In step S23, the manipulated variable calculation unit 114 obtains cluster numbers 142 from the status database 140 for a predetermined period from the current time. For example, the status calculation unit 112 obtains cluster numbers for the time period from the current time up to 30 minutes prior (current time, 10 minutes ago, 20 minutes ago, 30 minutes ago). In step S24, the manipulated variable calculation unit 114 uses the control manipulated variable calculation model 121 to calculate the control manipulated variable from the cluster number obtained in step S22.

[0055] In step S25, the display control unit 115 transmits the state name, current state, and the control operation amount calculated in step S24 from the state transition table 151 acquired in step S22 to the operation monitoring and control system 200. The current state is the state corresponding to the current (latest) state amount among the states (records) in the state transition table 151. The current state is the record that includes cluster number 142 of the latest record in the state database 140 in cluster number 152. In step S26, the control unit of the monitoring operation panel 203 displays a state progress display screen 510 (see Figure 8 below) on the monitoring operation panel 203, which includes the transition state and the control operation. The plant control unit 202 also performs feedback control based on the received control operation quantity.

[0056] Figure 8 is a screen configuration diagram of the state progress display screen 510 according to this embodiment. On the left side of the state progress display screen 510, the state names transmitted by the display control unit 115 are displayed in the order of the state transition table 151 (see step S22). The state names include the direction of change of the state quantity, the level of the state quantity compared to before the abnormality, and "recovery," which is in the process of returning to the state quantity before the abnormality.

[0057] In Figure 8, the state names up to the fifth state and the state names from the sixth state onward are displayed in a different form (for example, by different colors) (in Figure 8, the first five states are hatched). This is because the current state transmitted in step S25 is "Tower pressure: Recovering," indicating that the current state is "Tower pressure: Recovering" among the eight states that will transition sequentially.

[0058] On the right side of the status progress display screen 510, the control operation is displayed. The control operation amount calculated in step S24 is the control operation amount (negative bias value) that lowers the set value of the tower pressure, and is displayed as "Pressure set value: Lowering operation (-0.05)". The number in parentheses ("()") is the bias value. The status progress display screen 510 shows that the tower pressure is recovering, but has not yet fully recovered. Regarding control operations, it shows that a pressure reduction operation has been performed on the pressure setpoint and that this state is continuing. It also shows that although the elevated tower pressure is recovering, the raw material flow control valve opening has not yet reached the state where it can be set, so the pressure reduction operation is continuing and has not yet been reversed.

[0059] As explained above, the state of the controlled object (plant 300) displayed by the display control unit 115 includes the temporal changes in the state of the controlled object after the control operation variable has been controlled. Furthermore, the state of the controlled object displayed by the display control unit 115 indicates either the level of the state variable compared to the state variable before the abnormality of the controlled object was detected, the direction of change in the state variable, or the state being in the process of returning to the state variable before the abnormality of the controlled object was detected.

[0060] Returning to Figure 7, let's continue the explanation of the control process. In step S27, the display control unit 115 determines whether the plant 300 has settled. Settling means that the change in the state variable is below a predetermined value and is stable. If the plant has settled (step S27 → YES), the display control unit 115 proceeds to step S28; otherwise, if it has not settled (step S27 → NO), it returns to step S23. In step S28, the display control unit 115 transmits to the operation monitoring and control system 200 that the status progress display process has finished. Then, the control unit of the monitoring operation panel 203 terminates the display of the status progress display screen 510.

[0061] ≪Features of the control device≫ The control device 100 stores the state transitions of the plant 300 from an abnormal state to a stable state (see the state transition database 150 shown in Figure 6). When the control device 100 detects an abnormality, it identifies a state transition (state transition table 151) that starts with the abnormal state and transmits it to the operation monitoring and control system 200 (see step S25 shown in Figure 7). The control device 100 also calculates a control operation quantity from the state quantities at the current and past points in time and transmits it to the operation monitoring and control system 200 (see steps S23 to S25). This control operation quantity is calculated by referring to the control operation quantity calculation model 121, which has learned the control operations performed by a skilled operator, and is the same control operation as that performed by a skilled operator. The monitoring operation panel 203 of the operation monitoring and control system 200 displays a state progress display screen 510 (see Figure 8).

[0062] The operation monitoring and control system 200 controls the plant 300 based on the control operation values ​​transmitted by the control device 100, enabling a faster return to a steady state from an abnormal state. Furthermore, operators can refer to the state progress display screen 510 to understand the state transitions after an abnormality occurs and the control operations performed in each state. By understanding not only the current state but also the state transitions after control operations, operators can understand the rationale (reason) for performing control operations and learn the control operations of experienced operators.

[0063] <<Variation example: Control operation by the operator>> In the embodiment described above, the plant control unit 202 performs feedback control based on the control operation variable calculated by the control device 100 (see step S26 in Figure 7). The control operation variable may be displayed on the monitoring operation panel 203, and the operator may manually change the set value by referring to the displayed control operation variable. The control operation variable is calculated using a control operation variable calculation model 121 that has learned the control operations of a skilled operator, enabling a transition to a steady state in a short time.

[0064] By performing control operations in this manner, even inexperienced operators can perform control operations similar to those of experienced operators and set up plant 300 in a short amount of time. The status progress display screen 510 shows not only the current state of plant 300 but also the state transitions (time-series changes) after the control operation is performed. Therefore, operators can confirm what states plant 300 will go through to reach a steady state (settle) after the control operation is performed, and gain a basis for performing the control operation. Ultimately, operators can perform control operations smoothly and acquire the control operations of experienced operators.

[0065] ≪Torture: Trend graph display screen≫ The status progress display screen 510 (see Figure 8) includes the transitioning state of the plant 300 after a control operation has been performed. The monitoring operation panel 203 may also display a trend graph showing the changes in individual state variables. Figure 9 is a screen configuration diagram of a trend graph display screen 520 according to a modified example of this embodiment. The trend graph display screen 520 displays a graph of state variables from before the abnormality occurred to the present time. In the trend graph display screen 520 of Figure 9, the trend graph of the state variables after settling is displayed. The horizontal axis of the trend graph is time, and near the end, all state variables and set values ​​have reached steady values. It can be seen that the tower pressure and raw material supply flow rate have returned to the initial (before the abnormality occurred) state variables, but the raw material supply pressure remains high and the raw material flow rate control valve opening remains low. By looking at the trend graph, the operator can visually grasp the state transition up to the present time.

[0066] The trend graph displayed on the trend graph display screen 520 is a graph showing the measured state variables (sensor values). It may also be a graph that combines the measured state variables with the predicted values ​​of the state variables until the plant 300 settles down. Figure 10 is a screen configuration diagram of the trend graph display screen 520A according to a modified example of this embodiment. The solid line graph shows the measured state quantities, and the dotted line graph shows the predicted values ​​of the state quantities until the plant 300 settles down. The predicted values ​​of the state quantities can be obtained from the state quantity vectors included in the cluster corresponding to cluster number 152 included in the state transition table 151 identified in step S22 described in Figure 7. By looking at the trend graph, the operator can visually grasp the state transitions until a steady state is reached.

[0067] As described above, the state of the controlled object displayed by the display control unit 115 includes a graph showing the time-series change of the state variables. The graph showing the time-series change of the state variables also includes the measured values ​​and predicted values ​​of the state variables of the controlled object.

[0068] ≪Example of modified diagram: Trend display screen shown in the system diagram≫ The trend graph display screen 520 described above shows the trend (change tendency) of the state variables in a graph. The trend of the state variables may also be displayed on the system diagram of the plant 300. Figure 11 is a screen configuration diagram of the trend display screen 530 shown in the system diagram according to a modified example of this embodiment. In the trend display screen 530 shown in the system diagram, upward-right arrows, rightward arrows, and downward-right arrows are placed near the pressure sensors 312, 316 (see Figure 1), the flow sensor 314, and the control valve 313, so that the trend of each state variable can be seen. The upward-right arrow indicates an increase, the rightward arrow indicates recovery or stabilization, and the downward-right arrow indicates a decrease. By looking at the trend display screen 530 shown in the system diagram, the trend of the state variables can be seen visually.

[0069] As described above, the state of the controlled object displayed by the display control unit 115 includes a system diagram of the controlled object that shows either the state quantities of the equipment included in the controlled object or the direction of change of the state quantities.

[0070] <<Modification: Multiple transition states>> In the embodiment described above, there was only one state transition table 151 corresponding to the state of plant 300 at the time the abnormality was detected. However, it is conceivable that there may be multiple state transition tables 151, with the state at the time the abnormality was detected as the initial state (record), because the state transitions may branch along the way. In such cases, multiple transition states will be displayed on the monitoring operation panel 203.

[0071] Figure 12 is a screen configuration diagram of a modified state progress display screen 540 according to this embodiment. The state progress display screen 540 displays two transition states 541 and 542, and control operations 543 and 544 corresponding to transition states 541 and 542, respectively. The triangle 545 between transition states 541 and 542 indicates a branch in the state transition. In other words, the triangle 545 indicates that for transition states 541 and 542, the first state "raw material supply flow rate: abnormally rising" and the second state "in-column pressure: rising" are the same, but the third state branches into "raw material flow rate control valve opening: falling" and "raw material flow rate control valve opening: lower limit reached". In transition state 542, the rise in in-column pressure cannot be addressed by operating the raw material flow rate control valve alone, and the raw material flow rate control valve opening has reached its lower limit. In this case, it is determined that recovery of the in-column pressure is impossible, and the raw material supply pump is stopped as control operation 544.

[0072] In Figure 12, the state has progressed to the fifth state of transition state 541, "In-tower pressure: recovering," and has not branched to transition state 542, indicating that the state transition shown in transition state 541 is continuing. According to this status progression display screen 540, the operator can understand that multiple status transitions may occur and prepare appropriate responses in advance. As described above, the display control unit displays the changes in the multiple states over time.

[0073] <<Other variations>> Although several embodiments of the present invention have been described above, these embodiments are merely illustrative and do not limit the technical scope of the present invention. The present invention can take various other embodiments, and furthermore, various modifications such as omissions and substitutions can be made without departing from the spirit of the present invention. These embodiments and their variations are included in the scope and spirit of the invention as described herein, and are included in the scope of the invention and its equivalents as described in the claims. [Explanation of Symbols]

[0074] 100 Control device 111 State quantity receiving unit 112 State Calculation Unit 113 Anomaly detection unit 114 Operation amount calculation section 115 Display Control Unit 121 Control Variable Calculation Model 130 Plant State Variable Database 140 State Database 150 State Transition Database 151 State Transition Table (State Transitions) 160 Normal Status Database 203 Monitoring and Operation Panel (Display Unit) 300 plants (controlled items)

Claims

1. A state calculation unit that calculates the state of the controlled object based on the state variables of the controlled object, An operation variable calculation unit that calculates a control operation variable for controlling the controlled object based on the above state, The system includes a display control unit that displays the state of the controlled object and the control operation amount on a display, The state of the controlled object displayed by the display control unit includes the changes in the state of the controlled object over time after the control operation of the control variable. The system further includes an abnormality detection unit that detects abnormalities in the controlled object by referring to a normal state database that stores the normal range of the state quantities of the controlled object, The aforementioned manipulated variable calculation unit is: When an abnormality is detected in the controlled object, the control operation amount is calculated. The display control unit, The system refers to a state transition database that stores state transitions showing the changes in the state over time from the initial state, which is the state of the controlled object in which the abnormality was detected, to the steady state, which is the state in which the amount of change in the state quantity of the controlled object falls below a predetermined value. The abnormality detected by the abnormality detection unit is used to display the changes in the state included in the state transition over time, with the abnormality being the initial state. A control device characterized by the following features.

2. The display control unit, Display the changes in multiple states over time. The control device according to feature 1.

3. The state of the controlled object displayed by the display control unit is: This indicates the relative level of the state variable compared to the state variable before the abnormality of the controlled object was detected, the direction of change of the state variable, or the state variable in the process of returning to the state variable before the abnormality of the controlled object was detected. The control device according to feature 1.

4. The state of the controlled object displayed by the display control unit is: Includes a graph showing the time-series change of the aforementioned state variable. The control device according to feature 1.

5. The graph showing the time-series change of the aforementioned state variables is, The state variables of the controlled object include measured and predicted values. The control device according to feature 4.

6. The state of the controlled object displayed by the display control unit is: The system diagram of the controlled object includes either a state variable in the equipment included in the controlled object or the direction of change of said state variable. The control device according to feature 1.

7. The aforementioned manipulated variable calculation unit is: The control operation amount is calculated based on the state at the time of calculation of the control operation amount and the state at a time prior to the calculation. The control device according to feature 1.

8. The control device A step of calculating the state of the controlled object based on the state variables of the controlled object, A step of calculating a control operation variable for controlling the controlled object based on the aforementioned state, The steps include displaying the state of the controlled object and the control operation amount on a display device, The state of the controlled object displayed on the display unit includes the changes in the state of the controlled object over time after the control operation of the control variable. The steps include: detecting an abnormality in the controlled object by referring to a normal state database that stores the normal range of the state variables of the controlled object; When an abnormality in the controlled object is detected, the step of calculating the control operation amount is performed. The system refers to a state transition database that stores state transitions showing the changes in the state over time from the initial state, which is the state of the controlled object in which the abnormality was detected, to the steady state, which is the state in which the amount of change in the state quantity of the controlled object falls below a predetermined value. The steps include: displaying the changes over time in the states included in the state transition, with the aforementioned abnormality being the initial state; and executing the following steps. A control method characterized by the following:

Images