Air conditioning control device and program therefor
The dual-control air conditioning system addresses instability in predictive control by switching between PID and model predictive control based on energy and accuracy, ensuring stable and efficient temperature and humidity management.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- TRINITY IND CORP
- Filing Date
- 2022-03-17
- Publication Date
- 2026-05-11
AI Technical Summary
Model-based predictive control in air conditioning systems for painting facilities experiences instability due to modeling errors, leading to increased energy consumption and reduced precision in temperature and humidity control.
An air conditioning control device with dual control units - PID and model predictive control - that automatically switches between them based on energy consumption or control accuracy, using learning data collection and machine learning to maintain precise control and reduce energy use.
The system ensures stable and precise temperature and humidity control by dynamically switching between control methods, reducing energy consumption and maintaining continuous operation despite modeling errors.
Smart Images
Figure 0007856232000001 
Figure 0007856232000002 
Figure 0007856232000003
Abstract
Description
Technical Field
[0001] The present invention relates to an air-conditioning control device that controls the temperature and humidity of the intake outside air by operating a plurality of types of air-conditioning equipment, and a program therefor.
Background Art
[0002] Generally, a painting facility includes devices such as a painting booth that applies paint to an object to be painted such as an automobile body, and a drying furnace that dries the paint on the object that has passed through the painting booth. In such a painting facility, air whose temperature and humidity have been adjusted by an air conditioner for the painting booth is sent into the painting booth, and painting is performed. Further, the air conditioner for the painting booth is composed of devices such as a heating device, a cooling device, a humidifying device (washer), and a blower fan, and control is performed so that the conditioned air reaches the target temperature and humidity by operating these devices in combination (for example, see Patent Documents 1 and 2). Further, as control in this case, PID control has been conventionally used.
[0003] By the way, the air conditioner for the booth in an automobile painting booth requires strict temperature and humidity control in order to maintain the painting quality of the product. In recent years, as a new means for realizing strict temperature and humidity control, a model predictive control (MPC) that performs control while predicting future reactions has been proposed. Here, model predictive control is a control method that appropriately captures the dynamics of an air-conditioning device and uses a modeled prediction model. Therefore, according to model predictive control, it is considered that higher control performance can be realized than conventional PID control, and stable control results can be easily obtained.
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
Patent Document 2
[0005] However, model-based predictive control can experience errors between the predictive model and the actual response due to unexpected disturbances or changes over time. When such modeling errors occur, the control results become less stable compared to PID control, making it impossible to maintain precise temperature and humidity control. Consequently, energy consumption increases, and the goal of reducing energy consumption for air conditioning cannot be achieved.
[0006] The present invention has been made in view of the above-mentioned problems, and its objective is to provide an air conditioning control device and a program therefor that can reliably reduce the energy consumption required for air conditioning by maintaining precise temperature and humidity control. [Means for solving the problem]
[0007] To solve the above problems, the invention described in Means 1 is an air conditioning control device for controlling the amount of operation of an air conditioning device in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning devices, comprising: a first air conditioning control unit that controls the amount of operation of the air conditioning device by PID control; a second air conditioning control unit that controls the amount of operation of the air conditioning device by model predictive control based on a predictive model; a calculation and comparison unit that calculates and compares the energy required to bring the conditioned air to a target point by the PID control and the energy required to bring the conditioned air to a target point by the model predictive control; and a control switching unit that automatically switches to the air conditioning control unit of the first air conditioning control unit and the second air conditioning control unit that requires less energy to bring the conditioned air to a target point and performs the control. The learning data collection unit collects learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area. The gist of this invention is an air conditioning control device characterized by being equipped with the following features.
[0008] To solve the above problems, the invention described in means 2 is an air conditioning control device for controlling the amount of operation of an air conditioning device in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning devices, comprising: a first air conditioning control unit that controls the amount of operation of the air conditioning device by PID control; a second air conditioning control unit that controls the amount of operation of the air conditioning device by model predictive control based on a predictive model; a calculation and comparison unit that calculates and compares the deviation from the target value when the conditioned air reaches the target point by the PID control and the deviation from the target value when the conditioned air reaches the target point by the model predictive control; and a control switching unit that automatically switches to the air conditioning control unit of the first air conditioning control unit and the second air conditioning control unit that has a smaller deviation from the target value when the conditioned air reaches the target point and performs control. The learning data collection unit collects learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area. The gist of this invention is an air conditioning control device characterized by being equipped with the following features.
[0009] Therefore, according to the inventions described in means 1 and 2, since the system is equipped with two air conditioning control units, a first air conditioning control unit that performs control by PID control and a second air conditioning control unit that performs control by model predictive control, it is possible to select either air conditioning control unit to control the amount of operation of the air conditioning equipment. The calculation comparison unit calculates and compares the energy required to bring the conditioned air to the target point using PID control and the energy required to bring the conditioned air to the target point using model predictive control. Alternatively, the calculation comparison unit calculates and compares the deviation from the target value when bringing the conditioned air to the target point using PID control and the deviation from the target value when bringing the conditioned air to the target point using model predictive control. The control switching unit then automatically switches to the air conditioning control unit that consumes less energy based on the comparison result of the above energy and performs control. Alternatively, the control switching unit automatically switches to the air conditioning control unit with a smaller deviation from the target value (in other words, the air conditioning control unit with higher control accuracy) based on the comparison result of the above deviations and performs control. For example, if a modeling error occurs during the execution of model predictive control, the control result will become less stable compared to PID control, and the accuracy of the control will decrease, which is expected to increase the energy consumption required for air conditioning. In this case, the control switching unit automatically switches from the second air conditioning control unit to the first air conditioning control unit to execute PID control. Therefore, since the system is not affected by modeling errors and precise temperature and humidity control is maintained, the energy consumption required for air conditioning can be reliably reduced.
[0010] The invention described in means 3 is, in means 1 or 2, The aforementioned learning data collection unit, While the first air conditioning control unit is performing the PID control, it collects training data for further training of the prediction model. Ruko This is the gist of it.
[0011] Therefore, according to the invention described in means 3, even if the second air conditioning control unit cannot perform model predictive control due to the occurrence of modeling errors, the first air conditioning control unit performs PID control, and during this time, the learning data collection unit collects learning data. As a result, critical temperature and humidity control can be continuously performed without interruption. Furthermore, learning data for additional training can be efficiently collected.
[0012] The invention described in means 4 is characterized in that, in means 3, the machine learning unit further comprises a unit that creates a new prediction model based on the learning data collected by the learning data collection unit while the first air conditioning control unit is performing the PID control.
[0013] Therefore, according to the invention described in means 4, while the first air conditioning control unit is performing PID control, the machine learning unit creates a new predictive model that eliminates modeling errors. Thus, it is possible to prepare for the update work of replacing the old predictive model with the latest predictive model.
[0014] The gist of the invention described in means 5 is that, in means 1 or 2, the second air conditioning control unit further comprises a sequential learning unit that sequentially provides the control results as learning data for each control step while the second air conditioning control unit is performing the model predictive control, thereby continuously improving the error of the predictive model.
[0015] Therefore, according to the invention described in means 5, while the second air conditioning control unit is performing model predictive control, the sequential learning unit collects learning data for additional learning to continuously improve the error of the predictive model. This makes it easier for the second air conditioning control unit to continuously perform precise temperature and humidity control. Furthermore, it allows for efficient collection of learning data for additional learning.
[0016] The invention described in means 6 is characterized in that, in means 4 or 5, the control switching unit updates the old prediction model with the latest prediction model when the new prediction model is completed, and then automatically switches from the first air conditioning control unit to the second air conditioning control unit to perform control.
[0017] Accordingly, according to the invention described in means 6, once additional learning is completed and a new prediction model is finalized, the system automatically switches from the first air conditioning control unit to the second air conditioning control unit, and model prediction control is performed based on the updated latest prediction model that eliminates modeling errors. As a result, the system returns to model prediction control, which is more stable than PID control, and can continue to perform critical temperature and humidity control without interruption.
[0018] The invention described in means 7 is characterized in that, in any one of means 1 to 6, the air conditioning system is an air conditioner for a paint booth that includes a preheating device, a humidifying device, a cooling device, and a reheating device as the air conditioning equipment.
[0019] The invention described in means 8 is an air conditioning control device comprising an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, a first air conditioning control unit that controls the amount of operation of the air conditioning equipment by PID control, and a second air conditioning control unit that controls the amount of operation of the air conditioning equipment by model predictive control based on a predictive model, a calculation comparison step of calculating and comparing the energy required to bring the conditioned air to a target point by the PID control and the energy required to bring the conditioned air to a target point by the model predictive control, and a control switching step of automatically switching to the air conditioning control unit of the first air conditioning control unit and the second air conditioning control unit that requires less energy to bring the conditioned air to a target point and performing control. The learning data collection operation, which involves collecting learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area, is performed while the first air conditioning control unit is performing the PID control, and the learning data collection operation is performed to collect learning data for the prediction model. The gist of this invention is a program for an air conditioning control device that includes the following features.
[0020] The invention described in means 9 is an air conditioning control device comprising an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, a first air conditioning control unit that controls the amount of operation of the air conditioning equipment by PID control, and a second air conditioning control unit that controls the amount of operation of the air conditioning equipment by model predictive control based on a predictive model, a calculation comparison step of calculating and comparing the deviation from the target value when the conditioned air reaches the target point by the PID control and the deviation from the target value when the conditioned air reaches the target point by the model predictive control, and a control switching step of automatically switching to the air conditioning control unit of the first air conditioning control unit and the second air conditioning control unit that has a smaller deviation from the target value when the conditioned air reaches the target point and performing control. The learning data collection operation, which involves collecting learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area, is performed while the first air conditioning control unit is performing the PID control, and the learning data collection operation is performed to collect learning data for the prediction model. The gist of this invention is a program for an air conditioning control device that includes the following features.
[0021] Accordingly, according to the inventions described in means 8 and 9, in the calculation comparison step, the energy required to bring the conditioned air to the target point using PID control and the energy required to bring the conditioned air to the target point using model predictive control are calculated and compared. Alternatively, the deviation from the target value when bringing the conditioned air to the target point using PID control and the deviation from the target value when bringing the conditioned air to the target point using model predictive control are calculated and compared. In the subsequent control switching step, based on this comparison result, the system automatically switches to the air conditioning control unit that consumes less energy and performs the control. Alternatively, the system automatically switches to the air conditioning control unit with the smaller deviation from the target value (in other words, the air conditioning control unit with higher control accuracy) and performs the control. For example, if a modeling error occurs during the execution of model predictive control, the control result will become less stable compared to PID control, and the accuracy of the control will decrease, which is expected to increase the energy consumption required for air conditioning. In this case, the control switching unit automatically switches from the second air conditioning control unit to the first air conditioning control unit to execute PID control. Therefore, it is not affected by modeling errors, and precise temperature and humidity control is maintained, thus reliably reducing the energy consumption required for air conditioning. [Effects of the Invention]
[0022] As described in detail above, according to the inventions of Claims 1 to 9, an air-conditioning control device and a program therefor can be provided that can surely reduce the energy consumption required for air conditioning by continuously maintaining strict temperature and humidity control.
Brief Description of the Drawings
[0023] [Figure 1] A block diagram for explaining the air-conditioning control device of an embodiment embodying the present invention. [Figure 2] A block diagram more specifically showing the connection relationship between the air-conditioning control device of the embodiment and the air conditioner for a painting booth. [Figure 3] A flowchart for explaining the air-conditioning control method performed by the air-conditioning control device of the embodiment. [Figure 4] A psychrometric chart for explaining the data collection method in the air-conditioning control method performed by the air-conditioning control device of the embodiment. [Figure 5] A block diagram showing the connection relationship between the air-conditioning control device of another embodiment, the air conditioner for a painting booth, and a heat pump.
Mode for Carrying Out the Invention
[0024] Hereinafter, the air-conditioning control device in an air-conditioning system 11 of an embodiment embodying the present invention will be described in detail based on FIGS. 1 to 4.
[0025] The air-conditioning system 11 of the present embodiment shown in FIG. 1 is the air-conditioning system 11 for painting equipment as described above, and includes an air-conditioning device 21 that adjusts the temperature and humidity of the taken-in outside air using a plurality of types of air-conditioning equipment, and an air-conditioning control device 31 that controls the operation amount of the air-conditioning equipment. The air-conditioning device 21 in the present embodiment is an air conditioner 21 for a painting booth configured to include a plurality of types of air-conditioning equipment (preheating device, humidifying device, cooling device, reheating device, etc.).
[0026] As schematically shown in Figure 1, the paint booth air conditioner 21 in the air conditioning system 11 of this embodiment includes a preheating device 22, a humidifier 23, a cooling device 24, a reheating device 25, a first sensor 27a, a second sensor 27b, a third sensor 27c, and the like. On the other hand, the air conditioning control device 31 in the air conditioning system 11 of this embodiment includes an air supply target input unit 32, an air supply setting calculation unit 33, a first air conditioning control unit 42, a second air conditioning control unit 43, a calculation comparison unit 44, a control switching unit 45, a learning data collection unit 46, a machine learning unit 47, a system identification output unit 51, and the like. Figure 2 shows a block diagram to explain the more specific connection relationships of each element.
[0027] The paint booths to which the conditioned air generated by the paint equipment air conditioning system 11 is supplied are generally installed in areas on the workpiece transport line where paint is applied to workpieces. A paint booth comprises a painting chamber, an air supply chamber located above the painting chamber to supply downflow air (a constant direction from top to bottom) to the painting chamber, and an exhaust chamber located below the painting chamber to exhaust the air from the painting chamber. In the paint booth of this embodiment, the conditioned air discharged from the paint booth air conditioner 21 is supplied to the painting chamber from the air supply chamber in a downflow manner.
[0028] In the paint booth's painting chamber, the object to be painted is sprayed with paint mist from a painting machine (not shown). At this time, paint mist that is oversprayed and scattered from the painting machine is discharged from the painting chamber to the exhaust chamber by downflow conditioned air acting within the chamber. In the exhaust chamber, paint mist contained in the air is captured using circulating water from the booth, and the paint is recovered. The air discharged from the exhaust chamber is then released into the atmosphere by a fan.
[0029] As shown in Figures 1 and 2, the paint booth air conditioner 21 (air conditioning system) in this embodiment is composed of multiple types of air conditioning equipment. This paint booth air conditioner 21 is a device that takes in air from outside the device, adjusts it to a predetermined temperature (for example, around 23°C) and a predetermined humidity (for example, around 70% RH), and supplies it to the paint booth. Specifically, this paint booth air conditioner 21 includes a preheater 22 (preheating device), a washer 23 (humidifying device), a cooling coil 24 (cooling device), a reheater 25, and a blower fan 26.
[0030] The preheater 22 is a type of temperature control means for adjusting the temperature of the intake air, and is a device for heating the air to raise its temperature in advance. The washer 23 is a type of humidity control means for adjusting the humidity of the intake air, and is a device for increasing the humidity of the air by spraying water onto the air that has passed through the preheater 22. The cooling coil 24 is a type of temperature control means for adjusting the temperature of the intake air, and is a cooling device for cooling the air that has passed through the washer 23 to lower its temperature. The reheater 25 is a type of temperature control means for adjusting the temperature of the intake air, and is a reheating device for heating the air that has passed through the cooling coil 24 to raise its temperature again. The blower fan 26 is an air pumping device for pressurizing and sending temperature-controlled and humidity-controlled air (i.e., conditioned air) to the paint booth.
[0031] Sensing means are provided at multiple locations in the paint booth air conditioner 21. Specifically, the paint booth air conditioner 21 is equipped with a first sensor 27a for measuring temperature and humidity, a second sensor 27b for measuring temperature and humidity, and a third sensor 27c for measuring temperature. The first sensor 27a is for measuring the temperature and humidity of the outside air before air conditioning and is located near the outside air intake of the paint booth air conditioner 21. The second sensor 27b is for measuring the temperature and humidity of the outside air after air conditioning and is located on the outlet side of the blower fan 26 from which the conditioned air is sent out. The third sensor 27c is for measuring the temperature of the outside air after it has passed through the preheater 22 and is located upstream of the washer 23.
[0032] As shown in Figures 1 and 2, the air conditioning control device 31 for painting equipment in this embodiment is a device for controlling the operation amount of air conditioning equipment, and is composed of one or more well-known computers consisting of a CPU and memory means (ROM, RAM), etc.
[0033] The storage means in the air conditioning control device 31 stores a program for temperature and humidity control, and the CPU in the air conditioning control device 31 reads this program from the storage means and executes it sequentially. In addition to this program, the storage means also stores data related to a psychrometric chart (psychrometric chart table) that represents the state values of the air on a coordinate system. Incidentally, enthalpy increases as you move to the upper right of the psychrometric chart and decreases as you move to the lower left.
[0034] As shown in Figures 1 and 2, the first air conditioning control unit 42 in the air conditioning control device 31 controls the manipulated values of the air conditioning equipment using PID control. PID control (Proportional-Integral-Differential Control) is a type of feedback control that controls the input value using three elements: the deviation between the output value and the target value, its integral, and its derivative. The first air conditioning control unit 42 in this embodiment is a PID controller 42 and has the same number of PID loops as the number of controlled objects (specifically, four). The PID controller 42 and each air conditioning device (i.e., preheater 22, washer 23, cooling coil 24, reheater 25) are electrically connected via a control unit changeover switch 48, an adder 49, and a driver circuit (not shown). The PID controller 42 and each of the sensors 27a to 27c are also electrically connected. Therefore, when the PID controller 42 is connected to each air conditioning unit via the control unit changeover switch 48, the PID controller 42 outputs a drive control signal to each controlled object, thereby controlling the operation amount of each air conditioning unit using PID control. As a result, the outside temperature and humidity are adjusted to reach the target temperature and humidity. In addition, the temperature and humidity measurement results are input from each of the sensors 27a to 27c. Therefore, the PID controller 42 can perform feedback control based on these measurement results.
[0035] As shown in Figures 1 and 2, the second air conditioning control unit 43 in the air conditioning control device 31 controls the operation amount of the air conditioning equipment by Model Predictive Control (MPC) based on a prediction model 53. The second air conditioning control unit 43 in this embodiment is an MPC controller 43, and the MPC controller 43 and each air conditioning piece of equipment (i.e., preheater 22, washer 23, cooling coil 24, reheater 25) are electrically connected via a control unit changeover switch 48, an adder 49, and a driver circuit (not shown). The MPC controller 43 is also electrically connected to each of the sensors 27a to 27c. Therefore, when the MPC controller 43 is connected to each air conditioning piece of equipment by the control unit changeover switch 48, a drive control signal is output from the MPC controller 43 to each controlled object, and the operation amount of each air conditioning piece of equipment is controlled by the MPC. The MPC controller 43 is configured to include an optimizer, which calculates the optimal temperature and humidity control based on the prediction model. As a result, the ambient temperature and humidity are adjusted to reach the target temperature and humidity. In addition, the temperature and humidity measurement results are input from each of the sensors 27a to 27c. Therefore, the MPC controller 43 can perform feedback control based on these measurement results.
[0036] The air supply target input unit 32 is electrically connected to the PID controller 42 and the MPC controller 43 via the air supply setting calculation unit 33. The air supply target input unit 32 is for inputting target values for the temperature and humidity of the conditioned air to be supplied to the paint booth, and is configured to include means such as a keyboard or touch panel. The output signal of the air supply target input unit 32 is input to the air supply setting calculation unit 33.
[0037] The air supply setting calculation unit 33 and the sensors 27a to 27c are electrically connected. Therefore, the temperature and humidity measurements output from the sensors 27a to 27c are input to the air supply setting calculation unit 33. Based on the input target temperature and humidity values and the measured values, the air supply setting calculation unit 33 performs calculations to calculate the minimum enthalpy required to reach the target temperature and humidity. Based on the calculation results, the air supply setting calculation unit 33 sets target values for the operation amounts of each air conditioning device and outputs these target values to the PID controller 42 and the MPC controller 43.
[0038] The calculation and comparison unit 44 calculates and compares the energy required to bring the conditioned air to the target point using PID control and the energy required to bring the conditioned air to the target point using MPC. In this case, the energy required to bring the conditioned air to the target point using PID control is calculated based on the manipulated amount (control amount) of each air conditioning device determined by the PID controller 42, for example. The energy required to bring the conditioned air to the target point using MPC is calculated based on the manipulated amount (control amount) of each air conditioning device determined by the MPC controller 43, for example. The calculation and comparison unit 44 can also be understood by comparing the stability of PID control and the stability of MPC.
[0039] The control switching unit 45 automatically switches to the air conditioning control unit that requires less energy to get the conditioned air to the target point, among the PID controller 42 and the MPC controller 43, and performs the control. In other words, it automatically switches to the air conditioning control unit that can perform more stable temperature and humidity control, among the PID controller 42 and the MPC controller 43, and performs the control. Specifically, the control switching unit 45 is electrically connected to the calculation comparison unit 44 and operates based on the comparison results output from the calculation comparison unit 44. This control switching unit 45 is electrically connected to the control unit changeover switch 48. The control switching unit 45 connects one of the PID controller 42 and the MPC controller 43 to the respective air conditioning equipment by switching the control unit changeover switch 48. It should also be understood that this control switching unit 45 is a fail-safe unit that automatically switches to a relatively stable PID control when it detects undesirable behavior such as the occurrence and expansion of modeling errors during MPC execution.
[0040] The system identification output unit 51 is a part that randomly outputs step signals used for system identification. In the adder 49, the output signal from the system identification output unit 51 is added to the manipulated variable command signals from the PID controller 42 and the manipulated variable command signals from the MPC controller 43. The system identification output unit 51 is activated when random vibration is performed on the air conditioning equipment, which will be described later.
[0041] Furthermore, the air conditioning control device 31 includes a learning data collection unit 46 that collects training data for the predictive model used in the MPC of the paint booth air conditioner 21. The learning data collection unit 46 collects training data for additional training of the predictive model while the PID controller 42 is performing PID control. Specifically, as shown in Figure 2, the learning data collection unit 46 collects measured values (PV), controlled variables (or manipulated variables, MV), and actual results as training data and creates an actual results database. The program for collecting the learning data is stored in the storage means of the air conditioning control device 31. The CPU in the air conditioning control device 31 reads the program from the storage means as needed and executes it sequentially.
[0042] The learning data collection unit 46 collects data through the following steps: region setting, learning start point setting, state point movement, and data collection.
[0043] As shown in Figure 4, the region setting step involves defining the region R1 on the psychrometric chart for which a prediction model will be created, that is, the region to be targeted for temperature and humidity control (controlled region R1). This controlled region R1 can also be described as the region for which a high-quality prediction model is desired in order to achieve highly accurate temperature and humidity control.
[0044] In the learning start point setting step, multiple learning start points S1 to S9 are set within the set control target area R1, which will serve as the starting point for random vibration of the air conditioning equipment. Specifically, the number, position, and movement order of the learning start points S1 to S9 are also set. In this embodiment, nine points are set as learning start points S1 to S9 (see Figure 4). The number of learning start points S1 to S9 is preferably 10 or more. The positions of the learning start points S1 to S9 are not limited and can be set arbitrarily, for example, by setting them at positions spaced apart from each other on a psychrometric chart.
[0045] In the state point movement step, the air conditioning state point K1 is moved to the initial learning start point S1 by controlling the temperature and humidity of the air conditioning equipment using PID control.
[0046] In the data acquisition step, learning data is collected by randomly exciting the air conditioning equipment while moving the air conditioning air state point K1 between multiple learning start points S1 to S9 within the controlled region R1. Specifically, the system uses PID control to return the air conditioning air state point K1, which has been displaced from the learning start points S1 to S9, back to those learning start points S1 to S9 (see Figure 4). PID control is also used when moving the air conditioning air state point K1 from the current learning start point to the next learning start point.
[0047] While the PID controller 42 is performing PID control, the machine learning unit 47 creates a new prediction model 52 based on the training data collected by the training data collection unit 46. The newly created latest prediction model 52 is temporarily stored, for example, in a memory area within the machine learning unit 47. When the new prediction model 52 is completed, the control switching unit 45 replaces and updates the old prediction model 53 with the latest prediction model 52. After that, the control switching unit 45 automatically switches from the PID controller 42 to the MPC controller 43 and executes the MPC.
[0048] Next, the procedure for the temperature and humidity control method of this embodiment will be described based on the flowchart in Figure 3 and with reference to the psychrometric chart in Figure 4. Note that the procedure shown in this flowchart is just one example, and the temperature and humidity control method of this embodiment can, of course, be implemented using a different procedure.
[0049] In the temperature and humidity control method of this embodiment, step S110 is performed first. That is, the control switching unit 45 drives the control unit switching switch 48, connecting the MPC controller 43 to each air conditioning device. Then, the MPC controller 43 outputs a drive control signal to each controlled object, and the operating amount of each air conditioning device is controlled by the MPC. As a result, the temperature and humidity are adjusted by the temperature and humidity control based on the MPC so that the outside temperature and humidity reach the target temperature and humidity.
[0050] In the next step S120, the calculation and comparison unit 44 calculates the energy E1 required for the air-conditioning air to reach the target point by PID control and the energy E2 required for the air-conditioning air to reach the target point by MPC. In the next step S130, the calculated energies E1 and E2 are compared with each other. Specifically, it is determined whether the energy E1 required for the air-conditioning air to reach the target point by PID control is smaller than the energy E2 required for the air-conditioning air to reach the target point by MPC. If the determination result in step S130 is NO, that is, if E1≥E2, it is considered that MPC is performed based on a prediction model without modeling error. Therefore, the control result by MPC at the current time is in a more stable state than the control result by PID control. In this case, the process returns to step S110 and the temperature and humidity control based on MPC is continuously executed.
[0051] If the determination result in step S130 is YES, that is, if E1<E2, it is considered that MPC is performed based on a prediction model with modeling error and deterioration. Therefore, the control result by MPC at the current time is in a less stable state than the control result by PID control. In this case, the process proceeds to step S140, and the temperature and humidity control is automatically switched from MPC to PID control. That is, the control switching unit 45 drives the control unit switching switch 48 to switch the connection between the PID controller 43 and each air-conditioning device.
[0052] In the next step, S150, the learning data collection unit 46 is activated to collect learning data for the prediction model used for MPC control during PID control (see Figure 4). Specifically, first, a control target region R1 for creating the prediction model is set on the psychrometric chart. Next, multiple learning start points S1 to S9, which will serve as the starting point for random excitation of the air conditioning equipment, and their movement order are set within the set control target region R1. Next, the air conditioning air state point K1 is moved to the first learning start point S1 by PID control. Next, the system identification output unit 51 is activated and random excitation is performed starting from the learning start point S1. Then, control is executed to return the air conditioning air state point K1, which has been displaced from the learning start point S1, back to the learning start point S1 by PID control. Once data collection at the first learning start point S1 is complete, the air conditioning air state point K1 is moved from the current learning start point S1 to the next learning start point S2, and the same random excitation is performed. Afterward, this random excitation is performed sequentially from the learning start point S3 to S9.
[0053] In the next step, S160, the machine learning unit 47 activates to take in the training data collected by the training data collection unit 46 and performs machine learning to create the latest prediction model 52 based on that training data. Once the latest prediction model 52 is completed, the machine learning unit 47 temporarily stores the latest prediction model 52 in its own memory.
[0054] In the next step, S170, the control switching unit 45 activates to determine whether the latest prediction model 52 is complete. If the result of the determination in step S170 is NO, that is, if the latest prediction model 52 is not yet complete, the process returns to step S150 and continues with the collection of training data and machine learning. During this process, temperature and humidity control by PID control is maintained. If the result of the determination in step S170 is YES, that is, if the latest prediction model 52 is complete, the process moves to the next step, S180. The control switching unit 45 then activates to replace and update the old prediction model 53 with the latest prediction model 52. In the next step, S190, the process automatically switches from PID control to MPC and then returns to the initial step, S110. That is, the control switching unit 45 drives the control unit changeover switch 48 to switch the connection between the MPC controller 43 and each air conditioning unit. Then, the MPC controller 43 outputs a drive control signal to each controlled object, and each air conditioning unit returns to a state where it is controlled based on the input from the MPC.
[0055] Therefore, according to this embodiment, the following effects can be obtained.
[0056] (1) As described above, the air conditioning control device 31 of this embodiment includes a PID controller 42 which is a first air conditioning control unit, an MPC controller 43 which is a second air conditioning control unit, a calculation comparison unit 44, a control switching unit 45, etc. Therefore, it is possible to select either the PID controller 42 or the MPC controller 43 to control the amount of operation of the air conditioning equipment. The calculation comparison unit 44 calculates and compares the energy E1 required to bring the conditioned air to the target point using PID control and the energy E2 required to bring the conditioned air to the target point using MPC. Based on this comparison result, the control switching unit 45 automatically switches to the air conditioning control unit that consumes less energy and performs the control. For example, if a modeling error occurs during the execution of MPC, the control result will be less stable than that of PID control, and it is expected that the energy consumption required for air conditioning will increase. In this case, the control switching unit 45 automatically switches from the MPC controller 43 to the PID controller 42 and performs PID control. Therefore, since it is not affected by modeling errors and precise temperature and humidity control is maintained, the energy consumption required for air conditioning can be reliably reduced.
[0057] (2) In the air conditioning control device 31 of this embodiment, even if the MPC controller 43 is unable to execute MPC due to modeling errors, the PID controller 42 performs PID control, and during this time, the learning data collection unit 46 collects learning data. Therefore, critical temperature and humidity control can be performed continuously without interruption. In addition, learning data for additional learning can be collected efficiently.
[0058] (3) In the air conditioning control device 31 of this embodiment, while the PID controller 42 is performing PID control, the machine learning unit 47 creates a new prediction model 52 that eliminates modeling errors. Therefore, it is possible to be reliably prepared for the update work of replacing the old prediction model 52 with the latest prediction model 53.
[0059] (4) In the air conditioning control device 31 of this embodiment, once additional learning is completed and a new prediction model 52 is completed, the old prediction model 53 is promptly replaced and updated with the latest prediction model 52. Subsequently, the system automatically switches from the PID controller 42 to the MPC controller 43, and MPC is executed based on the updated latest prediction model 52, which has had modeling errors eliminated. As a result, the system returns to MPC, which is more stable than PID control, and can continue to execute critical temperature and humidity control without interruption.
[0060] Furthermore, each embodiment of the present invention may be modified as follows.
[0061] In the above embodiment, a paint booth air conditioner 21 was used, which was equipped with a preheater 22 (preheating device), a washer 23 (humidifying device), a cooling coil 24 (cooling device), a reheater 25, and a blower fan 26 as air conditioning equipment. However, the system is not limited to this, and different device configurations may be adopted. For example, the cooling coil 24 may be in two stages instead of one, or it may be omitted if unnecessary. The reheater 25 may also be omitted if unnecessary. In other words, the air conditioning device is not limited to one that has the functions of heating, humidifying, and cooling the incoming outside air, but may also have heating and humidifying functions but no cooling function, or have cooling and humidifying functions but no heating function.
[0062] In the above embodiment, the air conditioning system 11 of the present invention was embodied in an air conditioning system for painting equipment equipped with an air conditioner 21 for a painting booth, but it may of course also be embodied in an air conditioning system equipped with an air conditioning device for purposes other than painting booths.
[0063] In the above embodiment, the PID controller 42, which is the first air conditioning control unit, is configured to collect training data for additional training of the prediction model 53 while it is performing PID control, but it is not limited to this. For example, the latest external prediction model 53 created by another device of the same specifications may be incorporated and replaced with the old one. Conversely, the latest prediction model 53 created by the air conditioning control device 31 of this embodiment may not only be used in its own device, but may also be ported to another device of the same specifications and used therein.
[0064] In the above embodiment, an example was shown in which training data for additional training of the prediction model 53 is collected only while the first air conditioning control unit, the PID controller 42, is performing PID control, but the system is not limited to this. For example, the same training data collection may also be performed while the second air conditioning control unit, the MPC controller 43, is performing MPC. Specifically, for example, the system may be configured to include a sequential learning unit that continuously improves the error of the prediction model, and while the second air conditioning control unit is performing MPC, the control results may be sequentially provided to the sequential learning unit as training data for each control step. This configuration makes it easier for the second air conditioning control unit to continuously perform precise temperature and humidity control. It also allows for efficient collection of training data for additional training.
[0065] In the above embodiment, there were two air conditioning control units that controlled the amount of operation of the air conditioning equipment using two different control methods (PID control and MPC). The calculation and comparison unit 44 calculated and compared the energy required to bring the conditioned air to the target point for each of the two air conditioning control units, and the control switching unit 45 was configured to automatically switch to the air conditioning control unit that provided the less energy to perform the control. However, the system is not limited to this configuration. For example, there may be three or more air conditioning control units that controlled the amount of operation of the air conditioning equipment using three or more different control methods (PID control, MPC, and other control methods). The calculation and comparison unit 44 calculated and compared the energy required to bring the conditioned air to the target point for each of the three or more air conditioning control units, and the control switching unit 45 was configured to automatically switch to the air conditioning control unit that provided the least energy to perform the control. Alternatively, instead of a calculation and comparison unit 44 that calculates and compares energy for each of the three or more air conditioning control units, a calculation and comparison unit 44 that calculates and compares the stability of the control results for each of the three or more air conditioning control units may be used.
[0066] In the above embodiment, the calculation and comparison unit 44 calculates and compares the energy required to bring the conditioned air to the target point for each of the two air conditioning control units, and the control switching unit 45 is configured to automatically switch to the air conditioning control unit that uses less energy to perform control. However, the system is not limited to this configuration. For example, the calculation and comparison unit 44 may calculate and compare the deviation from the target value when bringing the conditioned air to the target point using PID control and the deviation from the target value when bringing the conditioned air to the target point using MPC. The control switching unit 45 may then automatically switch to the air conditioning control unit with the smaller deviation from the target value when bringing the conditioned air to the target point (in other words, the one with higher control accuracy) to perform control. With this configuration, even if a decrease in control accuracy is expected, the control switching unit will automatically switch the air conditioning control unit in advance, maintaining high control accuracy. Therefore, precise temperature and humidity control can be maintained, and the energy consumption required for air conditioning can be reliably reduced.
[0067] In the above embodiment, the calculation and comparison unit 44 performed energy calculation and comparison based on the temperature and humidity measurement results from three sensing means (first, second, and third sensors 27a, 27b, and 27c) provided on the paint booth air conditioner 21, and the PID controller 42 and MPC controller 43 performed feedback control, but the system is not limited to this. For example, measurement results from sensing means provided on equipment other than the paint booth air conditioner 21 may be used. In the air conditioning control device 31A of another embodiment shown in Figure 5, a heat source and chilled water are supplied to the paint booth air conditioner 21 from a heat pump (HP). The paint booth air conditioner 21 and the heat pump are connected by a first path 61 that supplies the heat source and a second path 62 that supplies chilled water. A fourth sensor 27d is provided on the first path 61 that supplies the heat source to the preheater 22. A fifth sensor 27e is provided on the first path 61 that supplies the heat source to the reheater 25. A sixth sensor 27f is provided on the second path 62 that supplies chilled water to the cooling coil 24. Examples of the fourth sensor 27d and fifth sensor 27e include a gas flow sensor, a steam flow sensor, and a hot water temperature flow sensor. Examples of the sixth sensor 27f include a chilled water flow sensor and a chilled water temperature sensor. In addition to the sensing information from sensors 27a to 27c, the calculation comparison unit 44 may perform an energy calculation comparison based on the sensing information from sensors 27d to 27f, and the PID controller 42 and MPC controller 43 may perform feedback control. By utilizing sensing information from energy-related sensors in this way, the energy calculation comparison by the calculation comparison unit 44 becomes more accurate, and consequently, the energy consumption required for air conditioning can be reduced more reliably. It is also possible to sense the operating power of the heat pump and use that sensing information for the above calculation comparison, etc.
[0068] Next, in addition to the technical ideas described in the claims, the technical ideas that can be grasped by the embodiments described above are listed below.
[0069] (1) An air conditioning control device for controlling the amount of operation of an air conditioning device in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning devices, comprising: a plurality of air conditioning control units that control the amount of operation of the air conditioning devices using different control methods; a calculation and comparison unit that calculates and compares the energy required to bring the conditioned air to a target point for each of the plurality of air conditioning control units; and a control switching unit that automatically switches to the air conditioning control unit among the plurality of air conditioning control units that requires the least amount of energy to bring the conditioned air to a target point and performs the control.
[0070] (2) An air conditioning control device for controlling the amount of operation of an air conditioning device in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning devices, comprising: a plurality of air conditioning control units that control the amount of operation of the air conditioning devices using different control methods; a calculation and comparison unit that calculates and compares the stability of the temperature and humidity control results; and a control switching unit that automatically switches to the air conditioning control unit with the highest stability among the plurality of air conditioning control units and performs the control.
[0071] (3) In any of the above means 1 to 9, the calculation comparison unit performs the calculation comparison of energy based on sensing information from the temperature and humidity sensor installed in the air conditioning device.
[0072] (4) In any of the above means 1 to 9, the calculation comparison unit performs the calculation comparison of energy based on sensing information from energy-related sensors installed on a path that connects the air conditioning device and the heat pump that supplies a heat source and chilled water to the air conditioning device in a flow path manner. [Explanation of symbols]
[0073] 21: Air conditioner for paint booths as an air conditioning system 22: Preheater as a preheating device for air conditioning equipment 23: Washers used as humidifiers in air conditioning equipment. 24: Cooling coil as a cooling device for air conditioning equipment. 25: Reheater as a reheating device for air conditioning equipment 31, 31A: Air conditioning control device 42: PID controller as the first air conditioning control unit 43: MPC controller as the second air conditioning control unit 44: Calculation and Comparison Section 45: Control switching section 46: Training Data Collection Department 47: Machine Learning Department 52: (Latest) Predictive Models 53: Predictive Models E1, E2: Energy
Claims
1. An air conditioning control device for controlling the amount of operation of air conditioning equipment in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, A first air conditioning control unit controls the amount of operation of the air conditioning equipment by PID control, A second air conditioning control unit controls the amount of operation of the air conditioning equipment by model predictive control based on a predictive model, A calculation and comparison unit calculates and compares the energy required to bring the conditioned air to the target point using the PID control and the energy required to bring the conditioned air to the target point using the model predictive control. A control switching unit that automatically switches to the air conditioning control unit that requires less energy to get the conditioned air to the target point, among the first air conditioning control unit and the second air conditioning control unit, The learning data collection unit collects learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area. An air conditioning control device characterized by being equipped with
2. An air conditioning control device for controlling the amount of operation of air conditioning equipment in an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, A first air conditioning control unit controls the amount of operation of the air conditioning equipment by PID control, A second air conditioning control unit controls the amount of operation of the air conditioning equipment by model predictive control based on a predictive model, A calculation and comparison unit calculates and compares the deviation from the target value when the conditioned air reaches the target point using the PID control and the deviation from the target value when the conditioned air reaches the target point using the model predictive control. A control switching unit that automatically switches to the air conditioning control unit that has a smaller deviation from the target value when the conditioned air reaches the target point, among the first air conditioning control unit and the second air conditioning control unit, and performs control accordingly. The learning data collection unit collects learning data for the prediction model by performing random excitation of the air conditioning equipment, which is an operation that returns the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area. An air conditioning control device characterized by being equipped with
3. The air conditioning control device according to claim 1 or 2, characterized in that the learning data collection unit collects learning data for additional training of the prediction model while the first air conditioning control unit is performing the PID control.
4. The air conditioning control device according to claim 3, further comprising a machine learning unit that creates a new prediction model based on the learning data collected by the learning data collection unit while the first air conditioning control unit is performing the PID control.
5. The air conditioning control device according to claim 1 or 2, further comprising a sequential learning unit that sequentially provides the control results as learning data for each control step while the second air conditioning control unit is performing the model predictive control, thereby continuously improving the error of the predictive model.
6. The air conditioning control device according to claim 4 or 5, characterized in that, when the new prediction model is completed, the control switching unit updates the old prediction model by replacing it with the latest prediction model, and then switches from the first air conditioning control unit to the second air conditioning control unit to perform control.
7. The air conditioning control device according to any one of claims 1 to 6, characterized in that the air conditioning device is an air conditioning device for painting equipment, and the air conditioning equipment is an air conditioner for a painting booth that includes a preheating device, a humidifying device, a cooling device, and a reheating device.
8. A program for operating an air conditioning control device comprising an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, a first air conditioning control unit that controls the operation amount of the air conditioning equipment by PID control, and a second air conditioning control unit that controls the operation amount of the air conditioning equipment by model predictive control based on a predictive model, A calculation and comparison step involves calculating and comparing the energy required to bring the conditioned air to the target point using the PID control and the energy required to bring the conditioned air to the target point using the model predictive control. A control switching step that automatically switches to the air conditioning control unit that requires less energy to get the conditioned air to the target point, among the first air conditioning control unit and the second air conditioning control unit, and performs control; The learning data collection operation, which involves collecting learning data for the prediction model by performing random excitation of the air conditioning equipment, which is the operation of returning the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area, is performed while the first air conditioning control unit is performing the PID control; A program for an air conditioning control device, characterized by including the following:
9. A program for operating an air conditioning control device comprising an air conditioning system that adjusts the temperature and humidity of incoming outside air using multiple types of air conditioning equipment, a first air conditioning control unit that controls the operation amount of the air conditioning equipment by PID control, and a second air conditioning control unit that controls the operation amount of the air conditioning equipment by model predictive control based on a predictive model, A calculation and comparison step involves calculating and comparing the deviation from the target value when the conditioned air reaches the target point using the PID control and the deviation from the target value when the conditioned air reaches the target point using the model predictive control. A control switching step that automatically switches to the air conditioning control unit that has a smaller deviation from the target value when the conditioned air reaches the target point, among the first air conditioning control unit and the second air conditioning control unit, and performs control; The learning data collection operation, which involves collecting learning data for the prediction model by performing random excitation of the air conditioning equipment, which is the operation of returning the air conditioning air condition point that has been displaced from a learning start point set within the prediction model creation area on the psychrometric chart to the learning start point using the PID control, at multiple learning start points set at different locations within the prediction model creation area, is performed while the first air conditioning control unit is performing the PID control; A program for an air conditioning control device, characterized by including the following: