Control support device, control support method, and program

Abstract

To make it possible to realize batch process control that can achieve the target control performance.SOLUTION: A control support device according to one aspect of the present disclosure is a control support device that supports control of an object that executes a batch process, and includes: an acquisition unit that acquires a control quantity output from the object and a target value for the control quantity; a prediction unit that calculates a first control quantity prediction time series that represents a time series of a future control quantity later than the control quantity, based on the control quantity, a model that represents a relation between an operation quantity and a control quantity of the object, and an operation pattern that represents the time series of the operation quantity; a selection unit that selects a best operation pattern from the plurality of operation patterns, based on the first control quantity prediction time series and the target value; and a correction unit that selects a best corrected operation pattern from among one or more corrected operation patterns obtained by correcting the best operation pattern by one or more correction methods, based on the target value.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a control assistance device, a control assistance method, and a program. [Background technology]

[0002] A characteristic of batch processes is that the same processing is repeated for each batch. For this reason, when identifying parameters for a batch process and controlling the process using those parameters, a technique is known in which information from the previous processing is also used for the next processing (for example, Patent Document 1). In addition, when controlling a batch process, the best batch operation pattern from the past, known as a golden batch, is also used. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 63-191202 Summary of the Invention [Problem to be solved by the invention]

[0004] However, due to various factors (such as the current state of the plant, disturbance conditions, changes in the characteristics of the equipment, etc.), the target control performance may not always be achieved by the golden batch.

[0005] The present disclosure has been made in consideration of the above points, and aims to make it possible to realize batch process control that can achieve target control performance. [Means for solving the problem]

[0006] A control assistance device according to one aspect of the present disclosure is a control assistance device that assists in controlling an object that executes a batch process, and includes: an acquisition unit that acquires a control quantity output from the object and a target value for the control quantity; a prediction unit that calculates a first control quantity prediction time series that represents a time series of a control quantity that is future than the control quantity, based on the control quantity, a model that represents the relationship between the operation quantity and control quantity of the object, and an operation pattern that represents the time series of the operation quantity; a selection unit that selects a best operation pattern from a plurality of the operation patterns, based on the first control quantity prediction time series and the target value; and a correction unit that selects a best corrected operation pattern from one or more corrected operation patterns that each correct the best operation pattern using one or more correction methods, based on the target value. [Effects of the Invention]

[0007] This makes it possible to realize batch process control that can achieve the target control performance. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 2 is a diagram illustrating an example of a hardware configuration of a control device according to the first embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a functional configuration of a control device according to the first embodiment. [Figure 3] FIG. 10 is a diagram showing an example of the first to nth batch operation patterns. [Figure 4] FIG. 10 is a diagram illustrating an example of the first to mth batch correction methods. [Figure 5] 10 is a flowchart illustrating an example of a best batch operation pattern selection process according to the first embodiment. [Figure 6] FIG. 10 is a diagram illustrating an example of a target deviation. [Figure 7] 10 is a flowchart illustrating an example of batch correction processing according to the first embodiment. [Figure 8] FIG. 10 is a diagram illustrating an example of a functional configuration of a control device according to a second embodiment. [Figure 9]FIG. 10 is a diagram illustrating an example of the (m+1)th to (m+m')th batch correction method. [Figure 10] 10 is a flowchart illustrating an example of batch correction processing according to the second embodiment. [Figure 11] FIG. 1 is a diagram schematically illustrating a control target according to an embodiment. [Figure 12] FIG. 4 is a diagram showing valve characteristics of an input valve to be controlled according to an embodiment. [Figure 13] FIG. 2 is a diagram illustrating a batch operation pattern, a CV value, a controlled variable, and a target value according to an embodiment. [Figure 14] FIG. 10 illustrates corrections to a best batch operation pattern according to one embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] First and second embodiments of the present invention will be described in detail below with reference to the drawings. In the first and second embodiments, a control device 10 will be described that realizes batch process control capable of achieving target control performance for a control object 20, such as a plant that executes a batch process. Specific examples of plants that execute batch processes include plants that execute manufacturing processes in the fields of steel, casting, food, semiconductors, etc. However, the control object 20 being a plant is just one example, and the control object 20 is not limited to a plant as long as it is a device, apparatus, facility, etc. that executes a batch process.

[0010] In a batch process, the same processing is repeated for each unit called a batch. Hereinafter, each batch-unit processing included in a batch process will be referred to as a "batch process." Also, a batch process is assumed to include at least one batch process.

[0011] In the following, the control amount of the control object 20 is defined as y, the target value for the control amount y at the end time of the batch processing is defined as r, and the manipulated variable for the control object 20 is defined as u. Furthermore, the variable representing time is defined as t, and the control amount at time t is defined as y(t) and the manipulated variable at time t is defined as u(t). Furthermore, the time is defined as a relative time from a certain predetermined reference time (e.g., "0:00"), etc.

[0012] [First embodiment] First, the first embodiment will be described. c Assuming the current time t c is assumed to be before the start of a certain batch process included in the batch process executed by the control target 20 and after the end of the batch process immediately before that batch process. In other words, the start time of a certain batch process is assumed to be t bgn , the end time is t end , the start time of the previous batch process is t bgn ', end time t end ', t end ' <t c <t bgn The start time of batch processing is also called the "batch start time," and the end time is also called the "batch end time."

[0013] <Example of Hardware Configuration of Control Device 10 According to First Embodiment> An example of the hardware configuration of a control device 10 according to the first embodiment is shown in Fig. 1. As shown in Fig. 1, the control device 10 according to the first embodiment includes an input device 101, a display device 102, an external I / F 103, a communication I / F 104, a RAM (Random Access Memory) 105, a ROM (Read Only Memory) 106, an auxiliary storage device 107, and a processor 108. Each of these pieces of hardware is connected to each other via a bus 109 so as to be able to communicate with each other.

[0014] The input device 101 is, for example, a keyboard, a mouse, a touch panel, a physical button, etc. The display device 102 is, for example, a display, a display panel, etc. Note that the control device 10 does not necessarily have to have at least one of the input device 101 and the display device 102, for example.

[0015] The external I / F 103 is an interface with an external device such as a recording medium 103a. Examples of the recording medium 103a include a CD (Compact Disc), a DVD (Digital Versatile Disk), an SD memory card (Secure Digital memory card), and a USB (Universal Serial Bus) memory card.

[0016] The communication I / F 104 is an interface through which the control device 10 communicates with other devices, apparatuses, terminals, etc. The RAM 105 is a volatile semiconductor memory (storage device) that temporarily stores programs and data. The ROM 106 is a non-volatile semiconductor memory (storage device) that can store programs and data even when the power is turned off. The auxiliary storage device 107 is a non-volatile storage device such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), or a flash memory. The processor 108 is, for example, one of various arithmetic devices such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit).

[0017] 1 is an example, and the hardware configuration of the control device 10 is not limited to this. For example, the control device 10 may have multiple auxiliary storage devices 107 or multiple processors 108, may not have some of the hardware shown in the figure, or may have various hardware other than the hardware shown in the figure.

[0018] <Example of functional configuration of the control device 10 according to the first embodiment> FIG. 2 shows an example of the functional configuration of the control device 10 according to the first embodiment. As shown in FIG. 2, the control device 10 according to the first embodiment includes a controlled variable / target value acquisition unit 201, a controlled variable prediction unit 202, a batch selection unit 203, a batch correction unit 204, a display control unit 205, and an output unit 206. These units are realized, for example, by a process in which one or more programs installed in the control device 10 are executed by the processor 108 or the like. The control device 10 according to the first embodiment also includes a batch operation pattern storage unit 207, a batch response model storage unit 208, and a batch correction method storage unit 209. These storage units are realized, for example, by a storage area of ​​the auxiliary storage device 107 or the like. However, at least one of these storage units may be realized, for example, by a storage area of ​​a storage device communicably connected to the control device 10.

[0019] The control variable / target value acquisition unit 201 acquires a given target value r and also acquires the current time t c Control amount y(t c ) is acquired from the control target 20. The target value r may be provided, for example, by an operation of an operator of the control target 20, or may be provided from a terminal used by the operator of the control target 20, or may be provided from other equipment, etc.

[0020] The controlled variable prediction unit 202 uses the batch response model stored in the batch response model storage unit 208 and the controlled variable y(t c ) and the batch operation pattern selected by the batch selection unit 203, a controlled variable prediction time series corresponding to the batch operation pattern is calculated. The batch operation pattern is time series data of the controlled variable u(t) when batch processing was executed in the past. The controlled variable prediction time series is the time series data of the controlled variable u(t) when batch processing was executed in the past at the current time t c From batch end time t endThe batch response model is time-series data of a predicted value of the controlled variable y up to a certain point (hereinafter also referred to as a "controlled variable predicted value"). Also, the batch response model is data that models the controlled variable y that is output from the controlled object 20 when the controlled variable u is input (in other words, data that models the relationship between the controlled variable u and the controlled variable y in the controlled object 20). The batch response model is expressed, for example, as a function or algorithm that takes at least the controlled variable u as input and outputs the controlled variable y.

[0021] The controlled variable prediction unit 202 also uses the batch response model stored in the batch response model storage unit 208 and the controlled variable y(t c ) and the corrected batch operation pattern calculated by the batch correction unit 204, a controlled variable predicted time series corresponding to the corrected batch operation pattern is calculated. The corrected batch operation pattern is time series data of the controlled variable u(t) obtained by correcting the best batch operation pattern (described later) selected by the batch selection unit 203 using a certain correction method.

[0022] The batch selection unit 203 selects a batch operation pattern to be used for calculating a control amount prediction time series from among the batch operation patterns stored in the batch operation pattern storage unit 207. When a control amount prediction time series has been calculated for all batch operation patterns stored in the batch operation pattern storage unit 207, the batch selection unit 203 selects the best batch operation pattern based on the target value r acquired by the control amount / target value acquisition unit 201 and each control amount prediction time series. The best batch operation pattern is the batch operation pattern from among the batch operation patterns stored in the batch operation pattern storage unit 207 that has obtained the best control amount prediction time series. Here, the quality of the control amount prediction time series is determined by the target value r and the batch end time t end For example, the control variable prediction time series with the smallest absolute value of the target deviation may be the best control variable prediction time series. endThe best controlled variable prediction time series may be the one in which the controlled variable prediction value does not exceed the target value r and the absolute value of the target deviation is smallest.

[0023] The batch correction unit 204 calculates a corrected batch operation pattern by correcting the best batch operation pattern using the batch correction method stored in the batch correction method storage unit 209 and the best batch operation pattern selected by the batch selection unit 203. The batch correction unit 204 also selects the best corrected batch operation pattern from among the corrected batch operation patterns as the best corrected batch operation pattern. The best corrected batch operation pattern is the corrected batch operation pattern from among the corrected batch operation patterns that has obtained the best controlled variable predicted time series.

[0024] The display control unit 205 causes the display device 102 to display the best corrected batch operation pattern selected by the batch corrector 204 and the controlled variable predicted time series corresponding to the best corrected batch operation pattern.

[0025] The output unit 206 outputs the best corrected batch operation pattern selected by the batch corrector 204 to the controlled object 20 .

[0026] The batch operation pattern storage unit 207 stores one or more batch operation patterns. The batch operation pattern storage unit 207 may store a batch operation pattern called a golden batch. Hereinafter, it is assumed that the batch operation pattern storage unit 207 stores n batch operation patterns (1st to nth), and the i-th batch operation pattern is referred to as {u i (t)|t∈[t bgn i -Δ1 i ,t end i +Δ2 i ]}. For simplicity, the i-th batch operation pattern will also be called "batch operation pattern i". Here, t bgn i is the batch start time of the batch process when the i-th batch operation pattern is performed, tend i is the batch end time of the batch process. i ≥ 0, Δ2 i ≧0 and t bgn i -Δ1 i ≦t <t bgn i For u i (t)=u i (t bgn i ), t end i <t≦t end i +Δ2 i For u i (t)=u i (t end i ) The batch operation pattern may be time series data of the operation amount u(t) when batch processing was actually performed during past operations, or may be time series data of the operation amount u(t) when batch processing was performed in a simulation or the like in the past. Specific examples of batch operation patterns will be described later.

[0027] The batch response model storage unit 208 stores the batch response model. A specific example of the batch response model will be described later.

[0028] The batch correction method storage unit 209 stores one or more batch correction methods. Hereinafter, it is assumed that the batch correction method storage unit 209 stores m batch correction methods, ie, the 1st to mth batch correction methods. For simplicity, the i-th batch correction method will also be referred to as "batch correction method i." Specific examples of the 1st to m-th batch correction methods will be described later.

[0029] <<Examples of 1st to nth batch operation patterns>> As an example, Figures 3(A) and 3(B) show specific examples of the first to third batch operation patterns when n = 3. Figure 3(A) shows batch operation pattern 1, Figure 3(B) shows batch operation pattern 2, and Figure 3(C) shows batch operation pattern 3, respectively.

[0030] ≪Specific Examples of the First to m-th Batch Correction Methods≫ As an example, specific examples of the first to second batch correction methods when m = 2 are shown in FIGS. 4(A) to 4(B).

[0031] The batch correction method 1 shown in FIG. 4(A) is a correction method that slides the batch operation pattern between the batch start time and the batch end time backward or forward for a predetermined time. For example, let the batch start time of the batch operation pattern to be corrected be t bgn,0 , and the batch end time be t end,0 . At this time, in the batch correction method 1 shown in FIG. 4(A), the batch start time is t bgn,1 = t bgn,0 + d1, and the batch end time is t end,1 = t end,0 + d1. Then, the batch operation pattern between the batch start time t bgn,0 and the batch end time t end,0 is slid backward or forward. Note that the operation amount at the time t satisfying t < t bgn,1 is the same as the operation amount at the batch start time t bgn,1 , and the operation amount at the time t satisfying t end,1 < t is the same as the operation amount at the batch end time t end,1 . Here, t bgn,1 is the batch start time of the corrected batch operation pattern, t end,1 is the batch end time of the corrected batch operation pattern, and d1 is a predetermined real number.

[0032] The batch correction method 2 shown in FIG. 4(B) is a correction method that shifts the batch end time backward or forward for a predetermined time. For example, let the batch start time of the batch operation pattern to be corrected be t bgn,0 , and the batch end time be t end,0 . At this time, in the batch correction method 2 shown in FIG. 4(B), the batch end time is t end,1 = t end,0 + d2. Note that the operation amount at the time t satisfying t end,1 < t is the same as the operation amount at the batch end time t end,1The same as the manipulated variable at t end,1 is the batch end time of the corrected batch operation pattern, and d2 is a predetermined real number.

[0033] Note that the batch correction methods shown in FIGS. 4(A) and 4(B) are all examples, and various other batch correction methods may be stored in the batch correction method storage unit 209.

[0034] <Best batch operation pattern selection process according to the first embodiment> The process of selecting the best batch operation pattern will be described with reference to FIG. 5. Hereinafter, for i=1, . . . , n, t c ∈[t bgn i -Δ1 i ,t bgn i ) is assumed. Also, given the target value r and the current time t c It is assumed that the control amount y up to is acquired by the control amount / target value acquisition unit 201.

[0035] The batch selection unit 203 initializes the value of a variable i used in the best batch operation pattern selection process to 1 (step S101). That is, the batch selection unit 203 sets i←1.

[0036] The batch selection unit 203 selects the ith batch operation pattern from the first to nth batch operation patterns stored in the batch operation pattern storage unit 207 (step S102).

[0037] The controlled variable prediction unit 202 uses the batch response model stored in the batch response model storage unit 208 and the controlled variable y(t c ) and the i-th batch operation pattern selected in step S102, a controlled variable prediction time series corresponding to the i-th batch operation pattern is calculated (step S103). The batch response model can be expressed, for example, by the following equation (1).

[0038]

number

[0039] However, the batch response model shown in the above formula (1) is just an example, and the batch response model can also be expressed by, for example, the following formula (2).

[0040]

number

[0041] As another example, the batch response model can be expressed by the following equation (3).

[0042]

number

[0043] The batch response model shown in the above equation (1), (2) or (3) is c The control amount y(t c ) as the base, the prediction target time t p This means that the controlled variable in is modeled as the integral of at least the manipulated variable u(t) transformed by the function f.

[0044] At this time, the control amount prediction unit 202 calculates u(t)=u i(t) is the batch response model shown in the above equation (1), (2), or (3) for each prediction target time t p ∈(t c ,t end i +Δ2 i ] predicted value of the controlled variable y p,1 i (t p ):=y(t p ) can be calculated. This allows us to calculate the control variable forecast time series {y p,1 i (t)|t∈(t c ,t end i +Δ2 i ]} is obtained.

[0045] The control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the control amount prediction time series calculated in the above step S103 and the target value r acquired by the control amount / target value acquisition unit 201 (step S104). That is, the control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the target value r and the batch end time t end i The predicted value of the controlled variable y p,1 i (t end i ) is the target deviation e i,1 Here, the target deviation e i,1 An example of the target deviation e is shown in Fig. 6. i,1 is the target value r and the batch end time t end i The predicted value of the controlled variable y p,1 i (t end i ) is defined as the difference between i,1 =ry p,1 i (t end i ) can be defined as

[0046] The batch selection unit 203 determines whether i>n-1 (step S105). In other words, the batch selection unit 203 determines whether the first to n-th batch operation patterns stored in the batch operation pattern storage unit 207 have been selected.

[0047] If it is not determined in step S105 above that i>n-1, the batch selection unit 203 adds 1 to the variable i (step S106) and returns to step S102 above. That is, the batch selection unit 203 sets i←i+1 and returns to step S102 above. As a result, steps S102 to S105 above are executed for all batch operation patterns, and n target deviations e 1,1 ,···,e n,1 is obtained.

[0048] On the other hand, if it is determined in step S105 that i>n-1, the batch selection unit 203 selects the best batch operation pattern from the first to n-th batch operation patterns (step S107). i,1 The batch operation pattern i corresponding to the controlled variable prediction time series with the smallest absolute value of y may be selected as the best batch operation pattern. p,1 i (t end i ) does not exceed the target value r, and the target deviation e i,1 The batch operation pattern i corresponding to the controlled variable forecast time series with the smallest absolute value of u may be selected as the best batch operation pattern. b (t)|t∈[t bgn b -Δ1 b ,t end b +Δ2 b ]} will be expressed as follows.

[0049] <Batch Correction Processing According to the First Embodiment> The batch correction process for calculating the best corrected batch operation pattern will be described with reference to Fig. 7. In the following, the best batch operation pattern {u b (t)|t∈[t bgn b -Δ1 b ,t end b +Δ2 b ]} is assumed to be obtained.

[0050] The batch correction unit 204 initializes the value of a variable i used in the batch correction process to 1 (step S201). That is, the batch correction unit 204 sets i←1.

[0051] The batch correction unit 204 selects the ith batch correction method from the first to mth batch correction methods stored in the batch correction method storage unit 209 (step S202).

[0052] The batch correction unit 204 calculates a corrected batch operation pattern i by correcting the best batch operation pattern by the i-th batch correction method selected in the above step S202 (step S203). Hereinafter, the corrected batch operation pattern i is referred to as {u b,i (t)|t∈[t bgn b,i -Δ1 b ,t end b,i +Δ2 b ]} where u b,i (t) is the operation amount included in the correction batch operation pattern i, t bgn b,i is the batch start time in the correction batch operation pattern i, t end b,i is the batch end time in the corrected batch operation pattern i.

[0053] The controlled variable prediction unit 202 uses the batch response model stored in the batch response model storage unit 208 and the controlled variable y(t c) and the corrected batch operation pattern i calculated in step S203, the control amount prediction unit 202 calculates a controlled variable prediction time series corresponding to the corrected batch operation pattern i (step S204). b,i (t) is the batch response model shown in the above equation (1), (2), or (3) for each prediction target time t p ∈(t c ,t end b,i +Δ2 b ] predicted value of the controlled variable y p,2 i (t p ):=y(t p ) can be calculated. This allows the controlled variable forecast time series {y p,2 i (t)|t∈(t c ,t end b,i +Δ2 b ]} is obtained.

[0054] The control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the control amount prediction time series calculated in the above step S204 and the target value r acquired by the control amount / target value acquisition unit 201 (step S205). That is, the control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the target value r and the batch end time t end b,i The predicted value of the controlled variable y p,2 i (t end b,i ) is the target deviation e i,2 It is calculated as follows: e i,2 =ry p,2 i (t end b,i ) can be defined as

[0055] The batch correction unit 204 determines whether i>m-1 (step S206). In other words, the batch correction unit 204 determines whether the first to m-th batch correction methods stored in the batch correction method storage unit 209 have been selected.

[0056] If it is not determined in step S206 that i>m-1, the batch correction unit 204 adds 1 to the variable i (step S207) and returns to step S202. That is, the batch correction unit 204 sets i←i+1 and returns to step S202. As a result, steps S202 to S206 are executed for the first to m-th batch correction methods, and m target deviations e 1,2 ,···,e m,2 is obtained.

[0057] On the other hand, if it is determined in step S206 that i>m-1, the batch correction unit 204 selects the best correction batch operation pattern from the correction batch operation patterns 1 to m (step S208). i,2 The correction batch operation pattern i corresponding to the control variable prediction time series with the smallest absolute value of y may be selected as the best correction batch operation pattern. p,2 i (t end b,i ) does not exceed the target value r, and the target deviation e i,2 The correction batch operation pattern i corresponding to the controlled variable prediction time series with the smallest absolute value of may be selected as the best correction batch operation pattern.

[0058] The display control unit 205 displays the best correction batch operation pattern and the corresponding controlled variable prediction time series on the display device 102. However, the display control unit 205 may also display the best correction batch operation pattern and the corresponding controlled variable prediction time series on a display provided in a terminal (e.g., a terminal used by an operator of the controlled object 20) communicably connected to the control device 10. This enables, for example, the operator of the controlled object 20 to know the best correction batch operation pattern, which is a batch operation pattern that can best achieve the target control performance of the controlled object 20, and the controlled variable prediction time series when the best correction batch operation pattern is performed.

[0059] The above-mentioned best correction batch operation pattern is output by the output unit 206 to the control object 20. This makes it possible to control the control object 20 using the best correction batch operation pattern. However, the output unit 206 may output to the control object 20 a batch operation pattern determined by an operator or the like with reference to the best correction batch operation pattern and the corresponding controlled variable predicted time series.

[0060] <Summary of the First Embodiment> As described above, according to the control device 10 of the first embodiment, a batch operation pattern including a past golden batch and a current time t c Control amount y(t c ) and a batch response model, an optimal batch operation pattern is selected and then corrected, thereby realizing or supporting batch process control that brings the controlled variable closer to the target value.

[0061] [Second embodiment] Next, a second embodiment will be described. In the second embodiment, t bgn ≦t c <t end During batch processing, the sampling period T c This will explain the case where the batch correction process is repeated for each batch. As a result, even if an unknown disturbance occurs or the error in the batch response model increases after the start of the batch, the best batch operation pattern can be corrected in real time, making it possible to achieve the target control performance.

[0062] In the second embodiment, differences from the first embodiment will be mainly described, and a description of components that may be the same as those in the first embodiment will be omitted.

[0063] <Example of functional configuration of the control device 10 according to the second embodiment> An example of the functional configuration of the control device 10 according to the second embodiment is shown in Fig. 8. As shown in Fig. 8, the control device 10 according to the second embodiment has a timer 210 in addition to the units described in the first embodiment. The timer 210 is realized, for example, by a process in which one or more programs installed in the control device 10 are executed by the processor 108 or the like.

[0064] After the batch process starts, the timer 210 starts the sampling period T c The batch correction unit 204 is operated every sampling period T c The expression "cycle" refers to the period (time width) during which the controlled variable y(t) is acquired by the controlled variable / target value acquiring unit 201, and its value is set in advance.

[0065] Furthermore, in addition to the m batch correction methods (1st to mth), the batch correction method storage unit 209 also stores m' batch correction methods (m+1th to m+m'th) that are used in the batch correction process executed after the start of the batch process. Specific examples of the (m+1st) to (m+m'th) batch correction methods will be described later.

[0066] However, the (m+1)th to (m+m')th batch correction methods may include all or part of the 1st to mth batch correction methods. In other words, the 1st to mth batch correction methods and the (m+1)th to (m+m')th batch correction methods may overlap all or part of them.

[0067] <Example of the m+1th to m+m'th batch correction method> As an example, specific examples of the (m+1)th to (m+2)th batch correction methods when m'=2 are shown in FIGS. 9(A) and 9(B).

[0068] The batch correction method m+1 shown in FIG. 9(A) is a correction method that slides the batch operation pattern between the current time and the batch end time back by a predetermined time (in other words, extends the operation amount at the current time by a predetermined time). For example, if the batch end time of the batch operation pattern to be corrected is set to t end,0Let it be so. At this time, in the batch correction method m + 1 shown in Fig. 9(A), the operation amount at the current time t c is extended to t1 = t c + d m+1 until, and the batch end time is t end,1 = t end,0 + d m+1 Let it be so. Note that the operation amount at the time t satisfying t end,1 < t is the same as the operation amount at the batch end time t end,1 . Here, t end,1 is the batch end time of the corrected batch operation pattern, and d m+1 is a predetermined real number.

[0069] The batch correction method m + 2 shown in Fig. 9(B) is a correction method for shifting the batch end time backward or forward by a predetermined time. For example, let the batch end time of the batch operation pattern to be corrected be t end,0 . At this time, in the batch correction method m + 2 shown in Fig. 9(B), the batch end time is t end,1 = t end,0 + d m+2 Let it be so. Note that the operation amount at the time t satisfying t end,1 < t is the same as the operation amount at the batch end time t end,1 . Here, t end,1 is the batch end time of the corrected batch operation pattern, and d m+2 is a predetermined real number. [[ID=4,2]]

[0070] Note that the batch correction methods shown in Figs. 9(A) to 9(B) are all examples, and various other batch correction methods may be stored in the batch correction method storage unit 209.

[0071] <Batch correction process according to the second embodiment> Sampling period T c The batch correction process for calculating the best correction batch operation pattern every time will be described while referring to Fig. 10. Hereinafter, the best batch operation pattern {u<0, b (t)|t ∈ [t bgn b -Δ1b ,t end b +Δ2 b ]} is obtained, and {u b (t)|t∈(t c ,t end b +Δ2 b ]} is called the target best batch operation pattern. That is, the target best batch operation pattern is a batch operation pattern that is made up of time series data of operation amounts at future times among the best batch operation patterns. Note that the following steps S301 to S308 are performed with a sampling period T c It is executed repeatedly every time.

[0072] The batch correction unit 204 initializes the value of a variable i used in the batch correction process to m+1 (step S301). That is, the batch correction unit 204 sets i←m+1.

[0073] The batch correction unit 204 selects the ith batch correction method from among the (m+1)th to (m+m')th batch correction methods stored in the batch correction method storage unit 209 (step S302).

[0074] The batch correction unit 204 calculates a corrected batch operation pattern i by correcting the target best batch operation pattern by the i-th batch correction method selected in the above step S302 (step S303). Hereinafter, the corrected batch operation pattern i is referred to as {u b,i (t)|t∈(t c ,t end b,i +Δ2 b ]} where u b,i (t) is the operation amount included in the correction batch operation pattern i, t end b,i is the batch end time in the corrected batch operation pattern i.

[0075] The controlled variable prediction unit 202 uses the batch response model stored in the batch response model storage unit 208 and the controlled variable y(t c) and the corrected batch operation pattern i calculated in step S303, the control amount prediction unit 202 calculates a controlled variable prediction time series corresponding to the corrected batch operation pattern i (step S304). b,i (t) is the batch response model shown in the above equation (1), (2), or (3) for each prediction target time t p ∈(t c ,t end b,i +Δ2 b ] predicted value of the controlled variable y p,2 i (t p ):=y(t p ) can be calculated. This allows the controlled variable forecast time series {y p,2 i (t)|t∈(t c ,t end b,i +Δ2 b ]} is obtained.

[0076] The control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the control amount prediction time series calculated in step S304 and the target value r acquired by the control amount / target value acquisition unit 201 (step S305). That is, the control amount prediction unit 202 calculates the target deviation corresponding to the control amount prediction time series based on the target value r and the batch end time t end b,i The predicted value of the controlled variable y p,2 i (t end b,i ) is the target deviation e i,2 Calculate as follows.

[0077] The batch correction unit 204 determines whether i>m+m'-1 (step S306). In other words, the batch correction unit 204 determines whether the (m+1)th to (m+m')th batch correction methods stored in the batch correction method storage unit 209 have been selected.

[0078] If it is not determined in step S306 that i>m+m'-1, the batch correction unit 204 adds 1 to the variable i (step S307) and returns to step S302. That is, the batch correction unit 204 sets i←i+1 and returns to step S302. As a result, steps S302 to S306 are executed for the m+1th to m+m'th batch correction methods, and m' number of target deviations e m+1,2 ,···,e m+m',2 is obtained.

[0079] On the other hand, if it is determined in step S306 that i>m+m'-1, the batch correction unit 204 selects the best correction batch operation pattern from the correction batch operation patterns m+1 to m+m' (step S308). i,2 The correction batch operation pattern i corresponding to the control variable prediction time series with the smallest absolute value of y may be selected as the best correction batch operation pattern. p,2 i (t end b,i ) does not exceed the target value r, and the target deviation e i,2 The correction batch operation pattern i corresponding to the controlled variable prediction time series with the smallest absolute value of may be selected as the best correction batch operation pattern.

[0080] The display control unit 205 displays the best correction batch operation pattern and the corresponding controlled variable prediction time series on the display device 102. However, the display control unit 205 may, for example, display the best correction batch operation pattern and the corresponding controlled variable prediction time series on a display provided in a terminal communicably connected to the control device 10. This enables, for example, an operator of the controlled object 20 to know the best correction batch operation pattern, which is a batch operation pattern that can best achieve the target control performance of the controlled object 20, and the controlled variable prediction time series when the best correction batch operation pattern is performed.

[0081] The above-mentioned best correction batch operation pattern is output by the output unit 206 to the control object 20. This makes it possible to control the control object 20 using the best correction batch operation pattern. However, the output unit 206 may output to the control object 20 a batch operation pattern determined by an operator or the like with reference to the best correction batch operation pattern and the corresponding controlled variable predicted time series.

[0082] <Summary of the second embodiment> As described above, according to the control device 10 according to the second embodiment, even after the start of the batch, the current time t c Control amount y(t c ), it is possible to realize or support batch process control that brings the controlled variable closer to the target value, even when, for example, an unknown disturbance occurs or the error of the batch response model becomes large.

[0083] [Example] An example of the control device 10 according to the first embodiment will be described below.

[0084] In this embodiment, the controlled object 20 shown in FIG. 11 is the target. The controlled object 20 shown in FIG. 11 is a system that stores water in a tank 500, and includes an input valve 501 that supplies water from the top and an output valve 502 that discharges water from the bottom. In this embodiment, the liquid level in the tank 500 (i.e., the height of the water from the bottom of the tank 500) is the controlled variable y, and the target liquid level is the target value r. The valve opening of the input valve 501 is the manipulated variable u. Furthermore, with respect to the manipulated variable u, it is assumed that there is an equal percentage characteristic shown in FIG. 12 between the valve opening and the actual flow rate of water. The CV value is a coefficient that represents the capacity of a valve to allow fluid to flow.

[0085] In this embodiment, the batch response model shown in the following equation (4) is used.

[0086]

number

[0087] It is assumed that two batch operation patterns are stored in the batch operation pattern storage unit 207: batch operation pattern 1 shown in Fig. 13(A) and batch operation pattern 2 shown in Fig. 13(B). In both of these two batch operation patterns, the batch start time is 10 and the batch end time is 30. Note that Figs. 13(A) and (B) also show the CV value of the flow rate according to the controlled variable y, the target value r, and the valve opening.

[0088] 4A is stored in the batch correction method storage unit 209. However, it is assumed that d1=7 and the entire batch operation pattern is slid back by d1.

[0089] At this time, the best batch operation pattern selection process shown in Fig. 5 was performed, and e 1,1 = 7.05, e 2,1 = 6.84. Therefore, batch operation pattern 2, which has the smallest target deviation, was selected as the best batch operation pattern.

[0090] Then, the batch correction process shown in Fig. 7 was performed, and the best batch operation pattern was corrected by batch correction method 1 as shown in Fig. 14. That is, the best batch operation pattern was shifted backward so that the batch start time was changed from 10 to 17 and the batch end time was changed from 30 to 37. As a result, the target deviation e, which is the difference between the predicted value of the controlled variable and the target value r at the batch end time of the best corrected batch operation pattern, 1,2 is e 1,2 =-0.16.

[0091] From the above, it can be seen that the absolute value of the target deviation is smaller in the best corrected batch operation pattern than in the best batch operation pattern, and batch process control can be realized in which the controlled variable approaches the target value more closely.

[0092] The present invention is not limited to the above-described embodiments specifically disclosed, and various modifications, changes, and combinations with known technologies are possible without departing from the scope of the claims. [Explanation of symbols]

[0093] 10 Control device 101 Input Device 102 Display device 103 External I / F 103a Recording media 104 Communication I / F 105 RAM 106 ROM 107 Auxiliary storage 108 processors 109 Bus 201 Controlled variable / target value acquisition unit 202 Control amount prediction unit 203 Batch Selection Section 204 Batch Correction Section 205 Display control unit 206 Output section 207 Batch operation pattern memory unit 208 Batch response model storage unit 209 Batch correction method memory section 210 Timer

Claims

1. A control assistance device that assists in controlling a target that executes a batch process, an acquisition unit that acquires a control amount output from the target and a target value for the control amount; a prediction unit that calculates a first controlled variable predicted time series that represents a controlled variable time series that is future than the controlled variable, based on the controlled variable, a model that represents a relationship between the controlled variable and the manipulated variable of the target, and an operation pattern that represents a time series of the manipulated variable; a selection unit that selects the best operation pattern from among the plurality of operation patterns based on the first controlled variable prediction time series and the target value; a correction unit that selects a best corrected operation pattern from one or more corrected operation patterns obtained by correcting the best operation pattern by one or more correction methods based on the target value; A control assistance device having:

2. The prediction unit calculating a second controlled variable prediction time series representing a controlled variable time series in the future relative to the controlled variable based on the controlled variable, the model, and the corrected operation pattern; The correction unit The control assistance device according to claim 1 , wherein the best corrective operation pattern is selected from the one or more corrective operation patterns based on the second controlled variable prediction time series and the target value.

3. The selection unit 2. The control assistance device according to claim 1, wherein the operation pattern in which the control amount at the end of a batch of the first controlled variable prediction time series is closest to the target value, or the operation pattern in which the control amount at the end of the batch is closest to the target value and the control amount at the end of the batch does not exceed the target value, is selected as the best operation pattern.

4. The correction unit 3. The control assistance device according to claim 2, wherein the corrective operation pattern in which the control amount at the end of a batch of the second controlled variable prediction time series is closest to the target value, or the corrective operation pattern in which the control amount at the end of the batch is closest to the target value and the control amount at the end of the batch does not exceed the target value, is selected as the best corrective operation pattern.

5. The control assistance device according to claim 2 or 4, further comprising a display control unit that displays the best corrective operation pattern and the second controlled variable predicted time series corresponding to the best corrective operation pattern.

6. The control assistance device according to claim 1 , further comprising an output unit that outputs the best corrective operation pattern to the target.

7. The one or more correction methods include: a correction method for sliding a time series of the manipulated variable from the batch start time to the batch end time of the best operation pattern forward or backward by a predetermined time; a correction method for shifting the batch end time of the best operation pattern forward or backward by a predetermined time; The control assistance device of claim 1 , comprising:

8. The model is 2. The control assistance device according to claim 1, wherein the control amount at each prediction target time is calculated by adding the control amount acquired by the acquisition unit and an integral value of a value obtained by converting at least the manipulated variable using a predetermined function.

9. The acquisition unit The control amount is acquired at every predetermined cycle, The prediction unit calculating the second controlled variable prediction time series based on the controlled variable, the model, and the corrected operation pattern for each period; The correction unit 3. The control assistance device according to claim 2, wherein a best corrected operation pattern is selected from one or more corrected operation patterns obtained by correcting, by the one or more correction methods, a target operation pattern among the best operation patterns that represents a time series of operation amounts future than a current time, based on the second controlled variable predicted time series and the target value.

10. The one or more correction methods include: a correction method for sliding a time series of manipulated variables from the current time to the batch end time of the best correction pattern forward or backward by a predetermined time; a correction method for shifting the batch end time of the best operation pattern forward or backward by a predetermined time; The control assistance device of claim 9, comprising:

11. A control assistance device that assists in controlling a target that executes a batch process, an acquisition step of acquiring a control amount output from the target and a target value for the control amount; a prediction step of calculating a first controlled variable predicted time series representing a controlled variable time series in the future than the controlled variable, based on the controlled variable, a model representing a relationship between the controlled variable and the manipulated variable of the target, and an operation pattern representing a time series of the manipulated variable; a selection step of selecting the best operation pattern from the plurality of operation patterns based on the first controlled variable prediction time series and the target value; a correction procedure of selecting a best corrected operation pattern from one or more corrected operation patterns obtained by correcting the best operation pattern by one or more correction methods based on the target value; A control assistance method for performing the above.

12. A control assistance device that assists in controlling a target that executes a batch process, an acquisition step of acquiring a control amount output from the target and a target value for the control amount; a prediction step of calculating a first controlled variable predicted time series representing a controlled variable time series in the future than the controlled variable, based on the controlled variable, a model representing a relationship between the controlled variable and the manipulated variable of the target, and an operation pattern representing a time series of the manipulated variable; a selection step of selecting the best operation pattern from the plurality of operation patterns based on the first controlled variable prediction time series and the target value; a correction procedure of selecting a best corrected operation pattern from one or more corrected operation patterns obtained by correcting the best operation pattern by one or more correction methods based on the target value; A program that executes the following.

Images