Calculation unit, calculation method, and program

Abstract

To provide an arithmetic unit capable of alleviating the workload concerning feedback control of a control-object apparatus.SOLUTION: An arithmetic unit 10 in accordance with the present disclosure includes an arithmetic block 13 that, when first control is implemented to output a functional output value FV, which depends on a measured value PV measured for a control-object apparatus 21, using a model M trained using learning data, calculates a control parameter CP, which is employed in second control of outputting a control output value MV, which is to be sent to the control-object apparatus 21 and depends on the measured value PV, to the control-object apparatus 21 through feedback control, on the basis of the measured value PV and a target value SV relevant to the measured value PV.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to an arithmetic device, an arithmetic method, and a program.

Background Art

[0002] Conventionally, techniques related to the control of devices are known. For example, Patent Document 1 discloses a control device that switches which of a first control unit and a second control unit controls a controlled device. The first control unit outputs an operation amount of the controlled device according to a measurement value by at least one of feedback control or feedforward control. The second control unit outputs an operation amount of the controlled device according to a measurement value using a model learned using learning data.

[0003] Patent Document 1 discloses, as an example, a hybrid control device that can use an AI (Artificial Intelligence) control function and a PID (Proportional-Integral-Differential) control function while switching according to the situation. Such a control device executes control by switching between an AI control function and a PID control function according to the desired controllability.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Conventional hybrid control systems can switch between AI control and PID control functions to perform control, but the control parameters for the PID control function must be investigated and set in advance. These control parameters can be obtained through manual adjustment by experts or through automatic adjustment using methods such as the limit sensitivity method and the step response method. However, these adjustments are time-consuming and involve a large workload.

[0006] This disclosure aims to provide a computing device, a computing method, and a program that can reduce the workload related to feedback control of controlled equipment. [Means for solving the problem]

[0007] In some embodiments of the computing device, a first control is being performed in which a functional output value corresponding to a measured value of a controlled device is output using a model learned with training data, and a second control is being performed in which a control output value corresponding to the measured value is output to the controlled device by feedback control, and a calculation unit is used to calculate control parameters to be used in the second control in which a control output value corresponding to the measured value is output to the controlled device by feedback control, based on the measured value and a target value for the measured value.

[0008] This allows the computing unit to reduce the workload related to feedback control of the controlled device. While the first control is being executed, the computing unit calculates the control parameters to be used for the second control based on the measured values ​​and target values ​​for those measured values. As a result, the computing unit, as a hybrid control device, can automatically calculate the control parameters during the first control. Therefore, the user does not need to pre-set the control parameters to be used for the second control.

[0009] In one embodiment, the arithmetic unit comprises a first control unit that outputs the function output value to a second control unit to execute the first control, and a second control unit that obtains the control parameters from the arithmetic unit to execute the second control, and the first control unit or the second control unit may control the controlled device.

[0010] This allows the arithmetic unit to perform various calculations related to the calculation of control parameters and various processes for controlling the controlled device, all on its own. Therefore, the convenience for users of the arithmetic unit is improved.

[0011] In one embodiment of the arithmetic device, the arithmetic unit may acquire the function output value based on the first control, and when the amount of change in the acquired function output value exceeds a threshold, it may start acquiring the measured value, which serves as the starting point for calculating the control parameter.

[0012] This allows the arithmetic unit to perform various calculations related to the calculation of control parameters at the appropriate timing when the function output value fluctuates significantly and the measured value asymptotically approaches the target value. Therefore, the arithmetic unit can suppress calculations at unnecessary timings and reduce the computational load.

[0013] In one embodiment of the arithmetic device, the arithmetic unit may calculate a plurality of parameters for calculating the control parameter based on the rate of change of the measured value per unit time.

[0014] This allows the arithmetic unit to calculate control parameters with greater accuracy. For example, the arithmetic unit can calculate control parameters with greater accuracy in cases like the one shown in Figure 3, where the measured value monotonically approaches the target value.

[0015] In one embodiment of the computing device, the control parameters may include parameters corresponding to at least one of proportional control, integral control, or differential control. This allows the computing device to perform feedback control by the second control unit as PID control. The computing device can control the controlled device to reduce the difference between the target value and the current measured value through PID control by the second control unit.

[0016] In some embodiments, the calculation method includes the step of calculating control parameters to be used in a second control that outputs a control output value to the controlled device corresponding to the measured value, based on the measured value and a target value for the measured value, when a first control is being performed in which a function output value corresponding to a measured value of the controlled device is output using a model learned with training data.

[0017] This allows the computing device executing the calculation method to reduce the workload related to feedback control of the controlled device. While the first control is being executed, the computing device calculates the control parameters to be used for the second control based on the measured values ​​and target values ​​for those measured values. As a result, the computing device, as a hybrid control device, can automatically calculate the control parameters during the first control. Therefore, the user does not need to pre-set the control parameters to be used for the second control.

[0018] In some embodiments, the program causes the computing unit to perform an operation that includes the step of calculating control parameters to be used in a second control, which outputs a control output value to the controlled device corresponding to the measured value, based on the measured value and a target value for the measured value, when a first control is being performed in which a function output value corresponding to a measured value of the controlled device is output using a model learned with training data, and the second control is performed in which a control output value corresponding to the measured value is output to the controlled device by feedback control, based on the measured value and a target value for the measured value.

[0019] This allows the computing unit to reduce the workload related to feedback control of the controlled device. While the first control is being executed, the computing unit calculates the control parameters to be used for the second control based on the measured values ​​and target values ​​for those measured values. As a result, the computing unit, as a hybrid control device, can automatically calculate the control parameters during the first control. Therefore, the user does not need to pre-set the control parameters to be used for the second control. [Effects of the Invention]

[0020] According to the present disclosure, it is possible to provide an arithmetic unit, an arithmetic method, and a program capable of reducing the workload related to feedback control of a controlled device.

Brief Description of the Drawings

[0021] [Figure 1] FIG. 1 is a block diagram showing an example of the configuration of a system including an arithmetic unit according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a functional block diagram showing an example of the configuration of the arithmetic unit in FIG. 1. [Figure 3] FIG. 3 is a schematic diagram for explaining an example of the calculation process by the arithmetic unit in FIG. 1. [Figure 4] FIG. 4 is a flowchart showing an example of the process executed by the arithmetic unit in FIG. 1.

Embodiments for Carrying Out the Invention

[0022] Hereinafter, an embodiment of the present disclosure will be mainly described with reference to the accompanying drawings.

[0023] FIG. 1 is a block diagram showing an example of the configuration of a system 1 including an arithmetic unit 10 according to an embodiment of the present disclosure. With reference to FIG. 1, an example of the configuration of a system 1 including an arithmetic unit 10 according to an embodiment will be mainly described. The system 1 has equipment 20 in addition to the arithmetic unit 10. The equipment 20 has at least one controlled device 21 and at least one sensor 22.

[0024] The equipment 20 is equipped with a device including at least one controlled device 21. For example, the equipment 20 may be a plant or a composite device combining a plurality of devices. In the present disclosure, the "plant" includes, for example, industrial plants such as chemical and bio plants, wellheads including gas fields and oil fields, and plants for managing and controlling the surrounding areas thereof. In addition, the plant may include plants for managing and controlling power generation such as hydraulic, thermal, and nuclear power, plants for managing and controlling environmental power generation such as solar and wind power, and plants for managing and controlling water supply and dams.

[0025] The controlled device 21 is an instrument, machine, or apparatus that controls, for example, at least one physical quantity in the process of the equipment 20. The controlled device 21 includes, for example, actuators such as valves, pumps, heaters, fans, motors, and switches. In this disclosure, “physical quantity” includes, for example, the temperature, pressure, flow rate, and pH of fluids, including gases and liquids, generated in the equipment 20 in which the controlled device 21 is located, as well as the degree of corrosion and vibration of the equipment 20. However, the physical quantity may also include state parameters, such as temperature and pressure, related to actuators, including valves, motors, and switches.

[0026] In one embodiment, for example, the equipment 20 is configured with a plurality of controlled devices 21. Each controlled device 21 may be of a different type, or at least two or more of them may be of the same type.

[0027] The controlled device 21 may be controlled externally via a wired or wireless connection through a network (not shown), or it may be controlled manually. The controlled device 21 is controlled by a computing device 10 which functions as a control device, as described later. If the system 1 is equipped with multiple controlled devices 21, the multiple controlled devices 21 may have relationships that allow them to be controlled in conjunction. For example, such relationships may include master-slave relationships and other relationships that are not controlled independently.

[0028] The controlled device 21 acquires the control output value MV output from the arithmetic unit 10 and receives control from the arithmetic unit 10 so that the current value of the physical quantity asymptotically approaches the target value SV that has been set in advance as a set value. For example, the controlled device 21 may receive feedback control that reduces the difference between the target value SV and the current value. The set target value SV may be a fixed value or may be changed as appropriate.

[0029] Sensor 22 includes a module capable of measuring physical quantities inside and outside the equipment 20. More specifically, sensor 22 includes a module that performs a measurement process on the physical quantity to be measured and obtains the measured value PV. Sensor 22 provides measurement data, including the measured value PV obtained by the measurement, to the computing device 10.

[0030] The measurement data includes physical quantities used by the computing unit 10 to control the controlled device 21. The measurement data may show the measured values ​​PV of physical quantities measured for the controlled device 21, for example, the output values ​​from the controlled device 21, or values ​​that change depending on the output values.

[0031] The computing device 10 may be any general-purpose electronic device such as a PC (Personal Computer) or a smartphone, or it may be another electronic device dedicated to system 1. However, it is not limited to these; the computing device 10 may be one or multiple server devices capable of communicating with each other.

[0032] In addition to executing various calculation processes described later, the arithmetic unit 10 also functions as a control device for controlling the controlled device 21. The arithmetic unit 10 switches between controlling the controlled device 21 by either the first control or the second control. In this disclosure, "first control" includes, for example, control that outputs a function output value FV corresponding to a measured value PV measured for the controlled device 21, using a model M learned using training data. The function output value FV corresponds to a control output value MV that is ultimately output from the arithmetic unit 10 as an manipulated variable for controlling the controlled device 21. "Second control" includes, for example, control that outputs a control output value MV to the controlled device 21 corresponding to a measured value PV measured for the controlled device 21, to the controlled device 21 by feedback control.

[0033] The arithmetic unit 10 includes a first control unit 11, a second control unit 12, an arithmetic unit 13, and a storage unit 14.

[0034] The first control unit 11 includes one or more processors. In this disclosure, “processor” includes, for example, a general-purpose processor or a dedicated processor specialized for a particular process. The first control unit 11 outputs a function output value FV to the second control unit 12 to execute the first control. More specifically, the first control unit 11 obtains a measured value PV measured for the controlled device 21 from the sensor 22 of the equipment 20. The first control unit 11 obtains a control target value SV for the measured value PV, for example, from user input. The first control unit 11 obtains a model M learned using training data.

[0035] Model M may be generated, for example, in the arithmetic unit 10, or in another device different from the arithmetic unit 10. Model M is generated by a learning process performed using learning data that includes measurement data including the measured value PV obtained by the sensor 22 and the control output value MV output from the second control unit 12. The generated model M takes the measured value PV as input and outputs a function output value FV.

[0036] The first control unit 11 outputs a function output value FV corresponding to the acquired measurement value PV and target value SV to the second control unit 12 using model M. The first control unit 11 provides an AI control function to the controlled device 21 by outputting the function output value FV as a control output value MV to the controlled device 21 via the second control unit 12. The first control unit 11 may also output a control output value MV corresponding to the measurement value PV for each of the multiple controlled devices 21, thereby providing an AI control function for each of the multiple controlled devices 21.

[0037] The first control unit 11 outputs a signal S to the second control unit 12 for switching between operation in auto mode and operation in manual mode. In this disclosure, "auto mode" includes, for example, a mode in which the second control unit 12 calculates and outputs a control output value MV to the controlled device 21 corresponding to the measured value PV when the second control unit 12 receives a measured value PV as input to the second control unit 12. Auto mode corresponds to second control. "Manual mode" includes, for example, a mode in which the second control unit 12 outputs a control output value MV corresponding to the control output value MV that the second control unit 12 should output when the second control unit 12 receives a function output value FV from the first control unit 11 to the second control unit 12. Manual mode corresponds to first control.

[0038] The second control unit 12 includes one or more processors. The second control unit 12 obtains control parameters CP from the calculation unit 13 and executes second control. In this disclosure, the control parameters CP include parameters corresponding to at least one of proportional control (P control), integral control (I control), or differential control (D control).

[0039] The second control unit 12 outputs a control output value MV to the controlled device 21 corresponding to the measured value PV, for example, by PID control based on the control parameter CP calculated by the calculation unit 13. In this way, the second control unit 12 drives the controlled device 21 with the control output value MV. The second control unit 12 provides a PID control function for the controlled device 21. The second control unit 12 may also output a control output value MV corresponding to the measured value PV for each of the multiple controlled devices 21, thereby providing a PID control function for each of the multiple controlled devices 21. The second control unit 12 provides such a PID control function in auto mode operation based on the signal S acquired from the first control unit 11.

[0040] The second control unit 12 acquires the measured value PV of the controlled device 21 from the sensor 22 of the equipment 20 in order to perform feedback control in auto mode. Similarly, the second control unit 12 acquires the control target value SV for the measured value PV, for example, from user input. The second control unit 12 acquires the control parameter CP calculated by the calculation unit 13 from the calculation unit 13.

[0041] The second control unit 12 calculates a control output value MV based on the acquired measured value PV and target value SV, and the control parameter CP calculated by the calculation unit 13. The second control unit 12 calculates a control output value MV that reduces the difference between the target value SV set by, for example, the user, and the current measured value PV.

[0042] On the other hand, when the second control unit 12 receives a signal S from the first control unit 11 to switch to manual mode operation, it contributes to the provision of AI control functions by the first control unit 11 in manual mode operation. In manual mode, the second control unit 12 receives a function output value FV from the first control unit 11 that corresponds to the control output value MV that the second control unit 12 should output. Based on the acquired function output value FV, the second control unit 12 outputs the corresponding control output value MV. As a result, the second control unit 12 outputs the control output value MV to the controlled device 21 in such a way as to reduce the difference between the target value SV and the current measured value PV.

[0043] The arithmetic unit 10 functions as a control device that switches between controlling the controlled device 21 by the first control unit 11 or the second control unit 12. The arithmetic unit 10 functions as a hybrid control device that can, for example, switch between the AI ​​control function by the first control unit 11 and the PID control function by the second control unit 12 depending on the situation. The arithmetic unit 10 switches between the AI ​​control function and the PID control function to perform control according to the desired controllability.

[0044] For example, the first control unit 11 may perform such switching control by switching the operating mode of the second control unit 12 to auto mode or manual mode. By setting the second control unit 12 to auto mode, the first control unit 11 causes the second control unit 12 to perform PID control functions and control the controlled device 21. By setting the second control unit 12 to manual mode, the first control unit 11 itself performs AI control functions and controls the controlled device 21.

[0045] The storage unit 14 includes any storage device such as an HDD (Hard Disk Drive), SSD (Solid State Drive), EEPROM (Electrically Erasable Programmable Read-Only Memory), ROM (Read-Only Memory), and RAM (Random Access Memory). The storage unit 14 stores information necessary to realize the operation of the arithmetic unit 10. The storage unit 14 stores information obtained through the operation of the arithmetic unit 10. For example, the storage unit 14 stores system programs, application programs, and various data obtained by any means such as communication.

[0046] The storage unit 14 may function as a main memory, auxiliary memory, or cache memory. The storage unit 14 is not limited to one built into the arithmetic unit 10, but may also be an external storage device connected by a digital input / output port such as USB (Universal Serial Bus).

[0047] The arithmetic unit 13 includes one or more processors. The processors included in the first control unit 11, the second control unit 12, and the arithmetic unit 13 may be different from each other, or they may be common in at least some respects.

[0048] When the first control is being performed by the first control unit 11, the calculation unit 13 obtains a function output value FV based on the first control by the first control unit 11 from the first control unit 11. When the amount of change in the function output value FV obtained from the first control unit 11 exceeds a threshold, the calculation unit 13 starts obtaining the measured value PV.

[0049] The calculation unit 13 acquires the measured value PV of the controlled device 21 from the sensor 22 of the equipment 20. Similarly, the calculation unit 13 acquires the control target value SV for the measured value PV, for example, from user input.

[0050] The calculation unit 13 calculates the control parameter CP to be used for the second control based on the measured value PV and the target value SV for the measured value PV when the first control is being executed by the first control unit 11. The calculation unit 13 calculates the control parameter CP during the operation of the AI ​​control function, including, for example, during startup and when the target value SV is changed. The calculation unit 13 outputs the calculated control parameter CP to the second control unit 12 as the control parameter CP when the second control is executed by the second control unit 12, and reflects it there.

[0051] Figure 2 is a functional block diagram showing an example of the configuration of the calculation unit 13 in Figure 1. Figure 3 is a schematic diagram illustrating an example of the calculation process performed by the calculation unit 13 in Figure 1. The graph in the upper half of Figure 3 shows the time change of the measured value PV as the first control unit 11 performs the first control and the measured value PV asymptotically approaches the target value SV. The graph in the lower half of Figure 3 shows the time change of the function output value FV output from the first control unit 11 when the first control unit 11 is performing the first control. As shown in the graph, the calculation unit 13 starts acquiring the measured value PV when the function output value FV rises sharply and the amount of change exceeds a threshold.

[0052] As shown in Figure 2, the calculation unit 13 includes a first calculation unit 131, a second calculation unit 132, a third calculation unit 133, a fourth calculation unit 134, and a control parameter calculation unit 135. As shown in Figure 3, the calculation unit 13 calculates a plurality of parameters for calculating the control parameter CP based on the rate of change per unit time of the measured value PV. In this disclosure, the "plural of parameters" include, for example, the maximum value of the slope R corresponding to the rate of change per unit time of the measured value PV, a settling value K, a dead time L, and a time constant T.

[0053] As shown in Figure 2, the input to the calculation unit 13 includes, for example, a target value SV, a function output value FV, and a measured value PV controlled by the AI ​​control function. The output from the calculation unit 13 includes, for example, three parameters as control parameters CP, corresponding to P control, I control, and D control, respectively.

[0054] In the calculation unit 13, based on the above input, multiple parameters are calculated in four blocks: the first calculation unit 131, the second calculation unit 132, the third calculation unit 133, and the fourth calculation unit 134. The first calculation unit 131 calculates the maximum value of the slope R corresponding to the rate of change of the measured value PV per unit time. The second calculation unit 132 calculates the settling value K. The third calculation unit 133 calculates the dead time L. The fourth calculation unit 134 calculates the time constant T.

[0055] Referring also to Figure 3, the first calculation unit 131 calculates the slope R per unit time from the acquired current measurement value PV and the previous measurement value PV from the previous unit time. In this disclosure, "unit time" includes, for example, 1 s. The slope R is expressed by, for example, the following equation (1). R = (Current measurement PV - Previous measurement PV) / 1 (1)

[0056] Based on the flowchart in Figure 4 described later, once the first calculation unit 131 calculates the maximum value of the slope R, the second calculation unit 132, the third calculation unit 133, and the fourth calculation unit 134 calculate the settling value K, the dead time L, and the time constant T, respectively.

[0057] When the maximum value of the slope R is calculated by the first calculation unit 131, the second calculation unit 132 calculates a preset target value SV as the set value K. The second calculation unit 132 calculates the target value SV as the set value K when the updating of the maximum value of the slope R stops according to the flowchart in Figure 4, which will be described later. That is, the second calculation unit 132 calculates the set value K using the following equation (2). Settling value K = Target value SV (2)

[0058] The third calculation unit 133 calculates the wasted time L based on the maximum value of the slope R calculated by the first calculation unit 131 using the following formula (3). L = (R × T) R -PV R +PV0) / R (3) Here, T R This represents the time when the maximum value of the slope R was obtained. PV R PV0 represents the measured value PV when the maximum value of the slope R is obtained. PV0 represents the measured value PV when the calculation unit 13 starts acquiring the measured value PV.

[0059] The fourth calculation unit 134 calculates the time constant T based on the maximum value of the slope R calculated by the first calculation unit 131 using the following equation (4). T=K / R (4) Here, K is the set value K calculated by the second calculation unit 132.

[0060] The control parameter calculation unit 135 obtains the maximum value of the slope R calculated by the first calculation unit 131, the settling value K calculated by the second calculation unit 132, the dead time L calculated by the third calculation unit 133, and the time constant T calculated by the fourth calculation unit 134. The control parameter calculation unit 135 calculates the control parameter CP based on the multiple parameters obtained. For example, the control parameter calculation unit 135 calculates the control parameter CP, which includes three parameters related to PID control, using the Ziegler-Nichols step response method or the CHR (Chien-Hrones-Reswick) method.

[0061] The control parameter calculation unit 135 outputs the calculated control parameter CP to the second control unit 12, which then uses it for feedback control such as PID control, preparing for the switch from the first control to the second control. After the switch to the second control, the second control unit 12 uses the control parameter CP calculated by the control parameter calculation unit 135 to execute, for example, PID control. The timing of the switch from the first control to the second control is, for example, time T in the graph showing the time change of the measured value PV in Figure 3. R This includes any time after the time required for the output of the control parameter CP has elapsed, and before the measured value PV matches the target value SV.

[0062] Figure 4 is a flowchart illustrating an example of the processing performed by the arithmetic unit 10 in Figure 1. In addition to Figure 4, Figure 3 will also be referred to as appropriate, and an example of the operation realized by the arithmetic unit 13 of the arithmetic unit 10 will be mainly explained.

[0063] In step S100, the calculation unit 13 obtains a function output value FV based on the first control performed by the first control unit 11 from the first control unit 11 when the first control is being performed by the first control unit 11.

[0064] In step S101, the calculation unit 13 determines whether the change in the function output value FV obtained in step S100 exceeds a threshold. If the calculation unit 13 determines that the change in the function output value FV exceeds the threshold, it executes the process in step S102. If the calculation unit 13 determines that the change in the function output value FV does not exceed the threshold, it executes the process in step S100 again.

[0065] In step S102, if the calculation unit 13 determines in step S101 that the amount of change in the function output value FV exceeds the threshold, it starts acquiring the measured value PV, using the point when the amount of change in the function output value FV exceeds the threshold as the starting point for calculating the control parameter CP. At this time, the calculation unit 13 stores the measured value PV0 at the time the acquisition of the measured value PV started as information in the storage unit 14.

[0066] In step S103, the first calculation unit 131 of the calculation unit 13 acquires the current measured value PV of the controlled device 21 from the sensor 22 of the equipment 20.

[0067] In step S104, the first calculation unit 131 of the calculation unit 13 calculates the slope R per unit time using equation (1) from the current measured value PV obtained in step S103 and the previous measured value PV from the previous unit time.

[0068] In step S105, the first calculation unit 131 of the arithmetic unit 13 determines whether the slope R calculated in step S104 has increased compared to the slope R calculated in the previous step as step S104 is repeated through steps S103 from step S107. For example, the first calculation unit 131 determines whether the slope R calculated in step S104 is greater than the maximum value of the slope R calculated in the previous step. If the first calculation unit 131 determines that the slope R has increased, it executes the process in step S106. If the first calculation unit 131 determines that the slope R has not increased, it executes the process in step S108.

[0069] In step S106, if the first calculation unit 131 of the calculation unit 13 determines that the slope R has increased in step S105, it updates the previous past slope R to the increased current value.

[0070] In step S107, the first calculation unit 131 of the calculation unit 13 stores the measured value PV, the time, and the slope R as information in the storage unit 14 when the slope R was updated in step S106.

[0071] In step S108, the second calculation unit 132, the third calculation unit 133, and the fourth calculation unit 134 of the calculation unit 13, if they determine that the slope R has not increased in step S105, calculate the settling value K, dead time L, and time constant T from among several parameters based on the above equations (2) to (4). For example, in a graph of measured values ​​PV that increases monotonically toward the target value SV, as shown in the upper half of Figure 3, if the slope R has not increased, it means that the inflection point of the graph is time T R This means that the maximum value of the slope R is obtained at this point.

[0072] The calculation unit 13 calculates the maximum value of the slope R stored in the storage unit 14 in step S107, and the time T R , and measured PV R In step S102, the arithmetic unit 13 calculates the dead time L from the above multiple parameters based on the measured value PV0 stored in the memory unit 14 and equation (3). The arithmetic unit 13 calculates the preset target value SV as the set value K using equation (2). In step S107, the arithmetic unit 13 calculates the time constant T using equation (4) with the maximum value of the slope R stored in the memory unit 14 and the calculated set value K.

[0073] In step S109, the calculation unit 13 calculates the control parameter CP based on the multiple parameters obtained up to step S108. When the first control is being executed, the calculation unit 13 calculates the control parameter CP to be used for the second control based on the measured value PV and the target value SV for the measured value PV.

[0074] In step S110, the calculation unit 13 outputs the control parameter CP calculated in step S109 to the second control unit 12.

[0075] According to the arithmetic unit 10 of the above embodiment, the workload related to feedback control of the controlled device 21 can be reduced. When the first control is being executed, the arithmetic unit 10 calculates the control parameter CP to be used for the second control based on the measured value PV and the target value SV for the measured value PV. As a result, the arithmetic unit 10, as a hybrid control device, can automatically calculate the control parameter CP during the first control. Therefore, the user does not need to set the control parameter CP to be used for the second control in advance.

[0076] As a result, users can easily perform tasks related to feedback control. Even without users having to adjust the control parameter CP through trial and error beforehand, the AI ​​control function allows the computing unit 10 to calculate the control parameter CP during startup control, and then switches to the PID control function when it approaches the target value SV, thereby enabling stable control. In a hybrid control system, users can adjust the control parameter CP of PID control, which typically requires experience and time, in a short time and with a small workload.

[0077] The arithmetic unit 10 has a first control unit 11 and a second control unit 12, and functions as a control device that switches between controlling the controlled device 21 using either the first control unit 11 or the second control unit 12. As a result, the arithmetic unit 10 can perform various calculation processes related to the calculation of the control parameter CP and various processes for controlling the controlled device 21 on its own. Therefore, the convenience for users using the arithmetic unit 10 is improved.

[0078] The arithmetic unit 10 acquires the function output value FV based on the first control, and when the change in the acquired function output value FV exceeds a threshold, it starts acquiring the measured value PV. This allows the arithmetic unit 10 to perform various calculation processes related to the calculation of the control parameter CP at an appropriate timing when the function output value FV fluctuates significantly and the measured value PV asymptotically approaches the target value SV. Therefore, the arithmetic unit 10 can suppress calculation processing at unnecessary timings and reduce the computational load.

[0079] The calculation unit 10 calculates several parameters for calculating the control parameter CP based on the rate of change of the measured value PV per unit time. This allows the calculation unit 10 to calculate the control parameter CP with greater accuracy. For example, the calculation unit 10 can calculate the control parameter CP with greater accuracy in cases such as the one shown in Figure 3, where the measured value PV asymptotically approaches the target value SV monotonically.

[0080] By including a control parameter CP that corresponds to at least one of proportional control, integral control, or differential control, the arithmetic unit 10 can perform feedback control by the second control unit 12 as PID control. The arithmetic unit 10 can control the controlled device 21 to reduce the difference between the target value SV and the current measured value PV through PID control by the second control unit 12.

[0081] System 1 switches between controlling the controlled device 21 using either the first control unit 11 or the second control unit 12 of the arithmetic unit 10. This allows System 1 to perform control effectively using the other control unit if one of them cannot perform control effectively. In addition, unlike when only the first control unit 11 controls the controlled device 21, System 1 can delegate part of the control to the second control unit 12. Therefore, System 1 can also simplify the learning of the model M used in the first control unit 11.

[0082] The computing device 10 can also be configured, for example, by equipping an existing PID controller with the above-mentioned AI control function and a calculation function related to the control parameter CP. This makes it easy for the computing device 10 to achieve both a significant improvement in the control characteristics related to the rise time of the measured value PV and stability in steady-state PID control compared to an existing PID controller.

[0083] While this disclosure has been described based on the drawings and embodiments, it should be noted that those skilled in the art can make various modifications and alterations based on this disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of this disclosure. For example, the functions included in each configuration or step can be rearranged in a logically consistent manner, and multiple configurations or steps can be combined into one or divided.

[0084] For example, a general-purpose electronic device such as a smartphone or computer can function as the arithmetic unit 10 according to the above-described embodiment. Specifically, a program describing the processing content that realizes each function of the arithmetic unit 10 according to the embodiment is stored in the memory of the electronic device, and the processor of the electronic device reads and executes the program. Therefore, this disclosure can also be realized as a program that can be executed by a processor.

[0085] Alternatively, the disclosure may also be implemented as a non-temporary computer-readable medium storing a program executable by one or more processors for causing a computing device 10 or the like to perform each function according to one embodiment. It should be understood that these are also included within the scope of the disclosure.

[0086] In the above embodiment, the arithmetic unit 10 was described as a control device that performs a switch between controlling the controlled device 21 by the first control unit 11 or the second control unit 12, but it is not limited to this. The arithmetic unit 10 may also function as a server device that has a first control unit 11 that performs first control and a second control unit 12 that obtains control parameters CP from the arithmetic unit 13 and performs second control, and is communicably connected to a control device that performs a switch between controlling the controlled device 21 by the first control unit 11 or the second control unit 12.

[0087] The arithmetic unit 10 may be configured as a different device from the hybrid control device having the first control unit 11 and the second control unit 12 described above. The arithmetic unit 10 may be configured separately from the hybrid control device, primarily as a device that performs various arithmetic processing by the arithmetic unit 13.

[0088] As a result, the processing load related to the hybrid control device is eliminated for the arithmetic unit 10, and only the computational load based on various calculation processes performed by the arithmetic unit 13 is incurred. Therefore, the arithmetic unit 10 can reduce the load associated with its operation.

[0089] In the above embodiment, the arithmetic unit 10 was described as starting to acquire the measured value PV when the change in the acquired function output value FV exceeds a threshold, but it is not limited to this. The arithmetic unit 10 may start acquiring the measured value PV based on conditions other than those based on the change in the function output value FV. For example, the arithmetic unit 10 may start acquiring the measured value PV when the absolute value of the acquired function output value FV exceeds a threshold. The arithmetic unit 10 may acquire the measured value PV continuously rather than starting to acquire the measured value PV at a specific timing.

[0090] In the above embodiment, the arithmetic unit 10 was described as calculating multiple parameters for calculating the control parameter CP based on the rate of change of the measured value PV per unit time, but it is not limited to this. The arithmetic unit 10 may calculate multiple parameters based on a calculation method other than the method based on the rate of change of the measured value PV per unit time. The arithmetic unit 10 may calculate the control parameter CP based on a calculation method other than the method based on multiple parameters.

[0091] In the above embodiment, the control parameter CP was described as including, but is not limited to, a parameter corresponding to at least one of proportional control, integral control, or differential control. The feedback control performed by the second control unit 12 of the arithmetic unit 10 may be a control other than PID control.

[0092] In the above embodiment, the first control unit 11 switches between the AI ​​control function and the PID control function by switching the operating mode of the second control unit 12 to auto mode or manual mode, but the embodiment is not limited to this. In addition to the first control unit 11 and the second control unit 12, the arithmetic unit 10 may have other components such as a switching unit that performs such switching control. Such switching control may be performed by the switching unit instead of the first control unit 11.

[0093] In the above embodiment, as shown in Figure 3, the measured value PV under the first control asymptotically increases monotonically towards the target value SV, but this is not the only possible embodiment. The measured value PV under the first control may asymptotically decrease monotonically towards the target value SV, or it may not change monotonically at all.

[0094] Some embodiments of the present disclosure are described below. However, it should be noted that the embodiments of the present disclosure are not limited to these. [Note 1] When a first control is being performed in which a functional output value corresponding to a measured value of the controlled device is output using a model trained with training data, a calculation unit calculates control parameters to be used in a second control that outputs a control output value to the controlled device corresponding to the measured value via feedback control, based on the measured value and a target value for the measured value. Equipped with, Computing device. [Note 2] The arithmetic device described in Appendix 1, A first control unit that outputs the function output value to a second control unit and executes the first control, The second control unit obtains the control parameters from the calculation unit and executes the second control, Equipped with, The first control unit or the second control unit controls the controlled device. Computing device. [Note 3] The arithmetic device described in Appendix 1 or 2, The calculation unit acquires the function output value based on the first control, and when the amount of change in the acquired function output value exceeds a threshold, it starts acquiring the measured value, using this as the starting point for calculating the control parameter. Computing device. [Note 4] A computing device as described in any one of the appendices 1 to 3, The calculation unit calculates a plurality of parameters for calculating the control parameters based on the rate of change of the measured value per unit time. Computing device. [Note 5] A computing device as described in any one of the appendices 1 to 4, The control parameters include parameters corresponding to at least one of proportional control, integral control, or differential control. Computing device. [Note 6] When a first control is being performed in which a functional output value corresponding to a measured value of a controlled device is output using a model trained with training data, the control parameters used in a second control, which outputs a control output value to the controlled device corresponding to the measured value via feedback control, are calculated based on the measured value and a target value for the measured value. Calculation method. [Note 7] In the computing unit, When a first control is being performed in which a functional output value corresponding to a measured value of the controlled device is output using a model trained with training data, the system is made to perform an operation that includes the step of calculating control parameters to be used in a second control, which outputs a control output value to the controlled device corresponding to the measured value, based on the measured value and a target value for the measured value, through feedback control. program. [Explanation of symbols]

[0095] 1 System 10 Arithmetic unit 11 First Control Unit 12 Second Control Unit 13 Arithmetic section 131 First Calculation Unit 132 Second Calculation Unit 133 Third Calculation Section 134 4th Calculation Section 135 Control parameter calculation unit 14 Storage section 20 Equipment 21. Controlled Devices 22 sensors CP control parameters FV Function Output Value K Set value L wasted time M Model MV control output value PV measurement R slope S signal SV target value T time constant

Claims

1. A first control unit switches between a first control that outputs a function output value corresponding to a control output value to the controlled device according to a measured value of the controlled device, using a model trained with training data, and a second control that outputs the control output value to the controlled device according to the measured value through feedback control. A calculation unit that calculates control parameters to be used for the second control based on the measured value and the target value for the measured value while the first control is being executed. Equipped with, Computing device.

2. The arithmetic device according to claim 1, The first control unit outputs the function output value to the second control unit and executes the first control, The second control unit obtains the control parameters from the calculation unit and executes the second control, Equipped with, The first control unit or the second control unit controls the controlled device. Computing device.

3. The computing device according to claim 1 or 2, The calculation unit acquires the function output value based on the first control, and when the amount of change in the acquired function output value exceeds a threshold, it starts acquiring the measured value, using this as the starting point for calculating the control parameter. Computing device.

4. The computing device according to claim 1 or 2, The calculation unit calculates a plurality of parameters for calculating the control parameters based on the rate of change of the measured value per unit time. Computing device.

5. The computing device according to claim 1 or 2, The control parameters include parameters corresponding to at least one of proportional control, integral control, or differential control. Computing device.

6. A step of switching between a first control, which outputs a function output value corresponding to a control output value to the controlled device according to a measured value of the controlled device, using a model trained with training data, and a second control, which outputs the control output value to the controlled device according to the measured value by feedback control, The first control is being performed, and the second control includes the step of calculating the control parameters to be used in the second control based on the measured value and the target value for the measured value, Calculation method.

7. In the computing unit, The operation includes the steps of switching between a first control, which outputs a function output value corresponding to a control output value to the controlled device according to a measured value of the controlled device, using a model trained with training data, and a second control, which outputs the control output value to the controlled device according to the measured value through feedback control, and, while the first control is being executed, calculating the control parameters to be used in the second control based on the measured value and a target value for the measured value. program.

Images