JP2018106561A - Heat exchange system, controller, and neural network construction method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a heat exchange system and the like which can reduce adverse effects on heat exchange due to changes in a physical value.SOLUTION: A heat exchange system comprises: an inverter 102 for adjusting a flow rate of a cooling medium; a PID control part 11 for controlling a feedback control which outputs an operation amount to the inverter 102 so that the deviation between a blow-out air temperature and a blow-out air temperature setting value is 0; and a neural network 15 to which input values including values indicating fluctuations in an in-take air temperature affected by heat exchange (for example, a current value, a first order differential value, and a second order differential value) are input, and which controls the inverter 102 based on the input value. The neural network 15 performs learning with the deviation or the operation amount set to 0.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a heat exchange system, a controller, and a construction method of a neural network.

  For example, Patent Document 1 includes an air conditioning control that includes a neural network that learns operating characteristics depending on a difference in refrigeration cycles, and a control unit that controls the operation of an air conditioner based on the learning result of the neural network. An apparatus is disclosed.

JP-A-5-10568

  In the air conditioning control device described above, temperature fluctuations (physical quantity fluctuations) of air (fluid) before heat exchange may adversely affect the air conditioner operation control (heat exchange).

  The present invention relates to a heat exchange system that reduces adverse effects on heat exchange due to physical quantity fluctuations, a controller that reduces adverse effects on heat exchange caused by physical quantity fluctuations, and a neural network that reduces adverse effects on heat exchange caused by physical quantity fluctuations. The purpose is to provide a construction method.

(1) In order to achieve the above object, a heat exchange system according to the first aspect of the present invention includes:
A heat exchange system (for example, the heat exchange system 100) that performs heat exchange using a heat medium (for example, a refrigerant),
An adjusting device (for example, the inverter 102) that controls the heat exchange by adjusting the flow rate of the heat medium;
Feedback control for outputting an operation amount to the adjusting device so that a deviation between a first physical quantity (for example, blown air temperature) that changes according to the flow rate and a target value (for example, blown air temperature set value) becomes zero. A feedback control unit (for example, PID control unit 11) to perform;
A value (for example, current value, first-order differential value) representing a variation of a second physical quantity (for example, intake air temperature) that affects the heat exchange (for example, control of the temperature of the fluid after heat exchange by the PID control unit 11). And an input value including a second-order differential value), and a machine learning unit (for example, a neural network 15 that outputs an operation amount) that controls the adjusting device based on the input value, the deviation or the operation A machine learning unit (for example, a neural network 15 that performs learning to reduce the deviation or the operation amount) that performs learning to reduce the deviation or the operation amount using a quantity as a teacher signal;
Is a heat exchange system.

(2) In the heat exchange system of (1) above,
The input value includes a first-order differential value and a second-order differential value of the time change of the second physical quantity,
You may do it.

(3) In the heat exchange system of (1) or (2) above,
The fluctuation of the second physical quantity is a fluctuation of the temperature of the fluid before the heat exchange (for example, the air whose temperature is changed by heat exchange).
You may do it.

(4) In order to achieve the above object, a controller according to the second aspect of the present invention provides:
A controller (e.g., controller 120) that controls an adjustment device (e.g., inverter 102) that controls the heat exchange by adjusting a flow rate of a heat medium used for heat exchange;
Feedback control for outputting an operation amount to the adjusting device so that a deviation between a first physical quantity (for example, blown air temperature) that changes according to the flow rate and a target value (for example, blown air temperature set value) becomes zero. A feedback control unit (for example, PID control unit 11) to perform;
A value (for example, current value, first-order differential value) representing a variation of a second physical quantity (for example, intake air temperature) that affects the heat exchange (for example, control of the temperature of the fluid after heat exchange by the PID control unit 11). And an input value including a second-order differential value), and a machine learning unit (for example, a neural network 15 that outputs an operation amount) that controls the adjusting device based on the input value, the deviation or the operation A machine learning unit (for example, a neural network 15 that performs learning to reduce the deviation or the operation amount) that performs learning to reduce the deviation or the operation amount using a quantity as a teacher signal;
It is a controller provided with.

(5) In order to achieve the above object, a neural network construction method according to the third aspect of the present invention includes:
A method for constructing a neural network (for example, a production method for the neural network 15) for controlling an adjustment device (for example, the inverter 102) that controls the heat exchange by adjusting a flow rate of a heat medium used for heat exchange. ,
In feedback control that outputs an operation amount to the adjustment device so that a deviation between a first physical quantity (for example, blown air temperature) that changes according to the flow rate and a target value (for example, blown air temperature set value) becomes zero. The step of causing the neural network (for example, the neural network 15 for learning to reduce the deviation or the operation amount) to perform learning to reduce the deviation or the operation amount using the deviation or the operation amount as a teacher signal is provided. ,
The neural network is a value (for example, a current value) representing a variation in a second physical quantity (for example, intake air temperature) that affects the heat exchange (for example, control of the temperature of the fluid after heat exchange by the PID control unit 11). An input value including a first-order differential value and a second-order differential value) is input, and the adjustment device is controlled based on the input value (for example, an operation amount is output).
It is a construction method of a neural network.

  According to the present invention, it is possible to reduce adverse effects on heat exchange due to physical quantity fluctuations.

It is a lineblock diagram of the heat exchange system concerning one embodiment of the present invention. It is an example of the block diagram which shows a part of controller of FIG. It is a figure which shows the structural example of a neural network. (A) It is a figure which shows the fluctuation transition (time change) of intake air temperature. (B) is a figure which shows transition of the air blowing temperature when the change transition of (A) is given to the heat exchange system and the neural network is not used. (C) is a figure which shows transition of the air blowing temperature when the fluctuation | variation transition of (A) is given to the heat exchange system, and a neural network is used.

[Heat exchange system 100]
The heat exchange system 100 according to an embodiment of the present invention is configured as a direct expansion type cooling device that cools (cools) a room to be cooled.

  As shown in FIG. 1, the heat exchange system 100 includes a circulation path L1, a compressor 101, an inverter 102, a condenser 103, a cooling water channel 104, a cooling water flow rate adjustment valve 105, an electronic expansion valve 106, The fan 107, the evaporator 108, the operation unit 112, the controller 120, the pressure sensors SP1 to SP2, and the temperature sensors ST1 to ST3 are provided.

  Moreover, the heat exchange system 100 is roughly configured by an outdoor unit Y and an indoor unit Z. The outdoor unit Y is equipped with a compressor 101, an inverter 102, and the like, and the indoor unit Z is equipped with a blower 107, an evaporator 108, and the like. The indoor unit Z includes a suction port Z1 that sucks in air around the indoor unit Z, and a blowout port Z2 that blows out air cooled by the indoor unit Z.

  The circulation path L1 is a path for circulating the refrigerant. In the middle of the circulation path L1, a compressor 101, a condenser 103, an electronic expansion valve 106, and an evaporator 108 are arranged.

  The compressor 101 compresses the refrigerant (steam) flowing through the circulation path L1 to a high temperature and a high pressure. The compressor 101 also has a function of circulating the refrigerant. The compressor 101 compresses the refrigerant with a motor.

  The inverter 102 adjusts the output of the compressor 101 (flow rate of refrigerant discharged from the compressor 101). The inverter 102 adjusts the output of the compressor 101 by adjusting the rotation speed of the motor of the compressor 101.

  The condenser 103 condenses (cools and liquefies) the refrigerant compressed by the compressor 101. A cooling water passage 104 through which cooling water flows passes through the condenser 103, and the refrigerant is condensed by the cooling water passing through the cooling water passage 104.

  The cooling water flow rate adjustment valve 105 is disposed in the middle of the cooling water channel 104 and adjusts the flow rate of the cooling water flowing through the cooling water channel 104 based on the opening degree thereof.

  A pressure sensor SP1 is provided in the vicinity of the outlet of the condenser 103 (the outlet of the condensed refrigerant). The pressure sensor SP1 converts the pressure of the refrigerant condensed by the condenser 103 (also referred to as condensation pressure) into an electrical signal, and supplies the converted electrical signal (a signal indicating the condensation pressure (pressure value)) to the controller 120. . In this way, the pressure sensor SP1 detects the condensation pressure, and supplies the detected condensation pressure (pressure value) to the controller 120.

  The electronic expansion valve 106 adjusts the flow rate of the refrigerant condensed by the condenser 103 according to the opening degree, and rapidly depressurizes the refrigerant. Due to this pressure drop, the temperature of the refrigerant also decreases.

  The blower 107 blows air from a room to be cooled to the evaporator 108.

  The evaporator 108 evaporates the refrigerant whose temperature and pressure are reduced by the electronic expansion valve 106. The evaporation of the refrigerant by the evaporator 108 takes heat from the air blown by the blower 107. That is, the evaporator 108 performs heat exchange between the blown air and the evaporating refrigerant, and cools the air.

  The blower 107 and the evaporator 108 are mounted on the indoor unit Z. By the operation of the blower 107, air near the suction port Z1 of the room to be cooled is sucked into the indoor unit Z from the suction port Z1, the sucked air is cooled by the evaporator 108, and the cooled air is blown out from the blower port Z2. It is blown out of the air and supplied to the room to be cooled. Thereby, the room to be cooled is cooled.

  A temperature sensor ST1 is provided in the vicinity of the suction port Z1. The temperature sensor ST1 converts a temperature in the vicinity of the suction port Z1 (also referred to as suction air temperature) into an electrical signal, and supplies the converted electrical signal (a signal indicating the suction air temperature (temperature value)) to the controller 120. In this way, the temperature sensor ST1 detects the intake air temperature, and supplies the detected intake air temperature (temperature value) to the controller 120. The suction air temperature when the blower 107 is operating is the temperature of the air sucked from the suction port Z1 (air before cooling).

  A temperature sensor ST2 is provided in the vicinity of the air outlet Z2. The temperature sensor ST2 converts a temperature in the vicinity of the blowout port Z2 (also referred to as a blown air temperature) into an electrical signal, and supplies the converted electrical signal (a signal indicating the blown air temperature (temperature value)) to the controller 120. In this manner, the temperature sensor ST2 detects the blown air temperature and supplies the detected blown air temperature (temperature value) to the controller 120. It can be said that the blown air temperature when the blower 107 is operating is the temperature of the air (air blown from the blowout port Z2) after being cooled by the evaporator 108.

  In the vicinity of the refrigerant outlet of the evaporator 108, a pressure sensor SP2 and a temperature sensor ST3 are provided. The refrigerant outlet is an outlet through which the refrigerant evaporated by the evaporator 108 exits the evaporator. The pressure sensor SP2 converts the pressure of the refrigerant evaporated by the evaporator 108 (also referred to as evaporation pressure) into an electrical signal, and supplies the converted electrical signal (a signal indicating the evaporation pressure (pressure value)) to the controller 120. In this way, the pressure sensor SP2 detects the evaporation pressure, and supplies the detected evaporation pressure (pressure value) to the controller 120. The temperature sensor ST3 converts the temperature of the refrigerant evaporated by the evaporator 108 (also referred to as refrigerant vapor temperature) into an electric signal, and supplies the converted electric signal (a signal indicating the refrigerant vapor temperature (temperature value)) to the controller 120. To do. The temperature sensor ST3 detects the refrigerant vapor temperature, and supplies the detected refrigerant vapor temperature (temperature value) to the controller 120.

  The refrigerant circulates on the circulation path L1 in the order of the compressor 101 → the condenser 103 → the electronic expansion valve 106 → the evaporator 108 → the compressor 101. An accumulator that stores the refrigerant may be provided in the middle of the circulation path L1.

  The operation unit 112 includes a remote controller that receives an operation from the user. For example, the user operates the operation unit 112 and inputs desired values for the condensation pressure, the blown air temperature, and the superheat degree in the heat exchange system 100. The superheat degree is the superheat degree of the refrigerant evaporated by the evaporator 108, and is calculated from the refrigerant vapor temperature-saturation temperature (the saturation temperature at the evaporating pressure of the refrigerant). Each of the inputted values is supplied to the controller 120.

  The controller 120 includes various computers such as a PLC (programmable logic controller). For example, the controller 120 includes a ROM (Read Only Memory) that stores programs and data, and a nonvolatile storage that can change (including addition and deletion) stored information (programs and data) such as a hard disk and a flash memory. Controller based on a device (hereinafter, simply referred to as a non-volatile storage device) and a program stored in the ROM or non-volatile storage device, and using data stored in the ROM or non-volatile storage device A CPU (Central Processing Unit) that actually executes the processing executed by 120 and a RAM (Random Access Memory) that is a main memory of the CPU are provided.

  The controller 120 stores each value supplied from the operation unit 112 in a predetermined storage area of the nonvolatile storage device. Thereby, each of the above values is set as a set value. In addition, each set value set is also called a condensing pressure set value, a blown air temperature set value, and a superheat degree set value, respectively. Each set value is used as a target value in PID control or the like described later. At least one of the condensation pressure and the degree of superheat may be set as a fixed value by the manufacturer of the heat exchange system 100.

  The controller 120 controls the inverter 102, the cooling water flow rate adjustment valve 105, and the electronic expansion valve 106, thereby controlling the condensing pressure, the blown air temperature, and the degree of superheat, and bringing these close to the set values (target values). (Including matching and maintaining). The condensation pressure is controlled by the opening degree (cooling water flow rate) of the cooling water flow rate adjustment valve 105. The flow rate of the refrigerant (air temperature) is controlled by the inverter 102. The degree of superheat is controlled by the opening degree of the electronic expansion valve 106 (the flow rate of the refrigerant).

  The controller 120 supplies, for example, an operation amount that can take a numerical value in a predetermined range to each of the inverter 102, the cooling water flow rate adjustment valve 105, and the electronic expansion valve 106 (hereinafter also referred to as a control target). Control the controlled object. The operation amount takes a common numerical range of 0 to 100 [%] for each control object, but the operation based on the operation value of the same numerical value basically differs depending on the control object.

  In the inverter 102, the frequency of the AC power to be output is controlled by the operation amount. The said frequency is controlled according to the operation amount 0-100. The inverter 102 outputs AC power of the minimum frequency (minimum frequency that can be output by the inverter 102) when the operation amount is 0 [%], and the maximum frequency (the inverter 102 is AC power at the maximum frequency that can be output is output. Therefore, as the operation amount increases, the frequency increases (the operation amount indicates a ratio to the maximum frequency). In addition, the rotation speed of the motor of the compressor 101 is controlled by the frequency of the AC power. Therefore, the output of the compressor 101 (as a result, the flow rate of refrigerant output from the compressor 101 and the blown air temperature) is controlled by the operation amount (as a result, the heat exchange in the evaporator 108 is controlled). )

  In each of the cooling water flow rate adjustment valve 105 and the electronic expansion valve 106, the opening degree is controlled by the operation amount. The manipulated variable indicates the percentage of opening. Specifically, when the operation amount is 0 [%], the opening degree is 0 (the valve is closed). When the operation amount is 100 [%], the opening degree is fully open. When the operation amount is 20 [%], the opening is an opening corresponding to 20% of full opening.

  For example, the controller 120 calculates the operation amount by PID (Proportional-Integral-Differential) based on the deviation between the condensing pressure set value and the condensing pressure (feedback value) detected by the pressure sensor SP2. An operation amount is supplied to the cooling water flow rate adjustment valve 105 (PID control). By such PID control, the opening degree of the cooling water flow rate adjustment valve 105 (in other words, the condensing pressure) is adjusted so that the condensing pressure approaches the condensing pressure setting value (target value) (so that the deviation becomes 0). Be controlled.

  For example, the controller 120 calculates the degree of superheat derived from the evaporation pressure and the refrigerant vapor temperature detected by the pressure sensor SP2 and the temperature sensor ST3. The controller 120 derives a saturation temperature (saturation temperature at the evaporation pressure) based on the evaporation pressure (may be obtained by calculation, or may be obtained by referring to a table that defines the relationship between the evaporation pressure and the saturation temperature). Good), a value obtained by subtracting the saturation temperature from the refrigerant vapor temperature is acquired as the degree of superheat (the degree of superheat is fed back). The controller 120 calculates an operation amount by PID based on the deviation between the acquired superheat degree and the superheat degree set value, and supplies the calculated operation amount to the electronic expansion valve 106 (PID control). By such PID control, the opening degree (in other words, the degree of superheat) of the electronic expansion valve 106 is controlled so that the degree of superheat approaches the superheat degree set value (target value) (so that the deviation becomes 0). The

[Controller 120 for controlling blown air temperature (inverter 102)]
As shown in FIG. 2, the controller 120 that controls the inverter 102 includes a PID control unit 11, a fluctuation acquisition unit 13, and a neural network 15. For example, the configuration of FIG. 2 is realized by the PLC or the like operating as the PID control unit 11, the fluctuation acquisition unit 13, and the neural network 15.

  A deviation between the set value of the blown air temperature and the value (feedback value) of the blown air temperature detected by the temperature sensor ST2 is input to the PID control unit 11, and the PID control unit 11 is based on the input deviation. , The operation amount is calculated based on the PID, and the calculated operation amount is output to the inverter 102. In this way, the PID control unit 11 controls the inverter 102 so that the blown air temperature approaches the blown air temperature set value (target value) (so that the deviation becomes 0).

  An electric signal (a signal indicating the intake air temperature, here an analog signal) from the temperature sensor ST1 is constantly input to the fluctuation acquisition unit 13. The fluctuation acquisition unit 13 removes noise from the electric signal using, for example, a low-pass filter, converts the electric signal into a digital signal by sampling or the like, and a value indicating the fluctuation of the intake air temperature based on the converted digital signal. Is obtained (calculated here). In this embodiment, as the value indicating the fluctuation of the intake air temperature, the current value of the intake air temperature (the value of the intake air temperature input this time) and the time change of the intake air temperature (the value indicating the past intake air temperature). Are obtained by differentiating a first-order differential value (temperature change speed) and a second-order differential value (temperature change acceleration) obtained by differentiating (which is held in a predetermined storage unit). When the first-order differential value is Δt (time) = 1 (seconds), the first-order differential value is the difference (change amount) between the intake air temperature input last time and the intake air temperature input this time. The fluctuation of the intake air temperature acquired by the fluctuation acquisition unit 13 is output to the neural network 15. A value indicating the fluctuation of the intake air temperature may be acquired by an analog circuit (for example, a differentiation circuit or the like), digitally converted, and input to the neural network 15.

  In this embodiment, various sensors such as the temperature sensor ST2 supply the detected physical quantity (temperature or pressure) to the controller 120 in the form of an analog signal (analog / digital conversion is performed by the controller 120). The controller 120 may be supplied in the form of a digital signal.

As shown in FIG. 3, the neural network 15 has an input layer, an intermediate layer, and an output layer. The input value to the input layer of the neural network 15 is x _i , the connection load between the input layer and the intermediate layer is w _ij , the connection load between the intermediate layer and the output layer is w _jk , and the input to the element of the intermediate layer is s _j Then, the input s _j to the intermediate layer element can be expressed by the following mathematical formula (1), and the output q _j from the intermediate layer to the output layer can be expressed by the following mathematical formulas (2) and (3). The output y _k of 15 can be expressed by the following mathematical formula (4). Here, the subscripts i, j, and k are element numbers of the input layer, the intermediate layer, and the output layer, respectively, and n and m are the number of inputs and the number of intermediate elements, respectively. In the following formula, x ₀ = 1, s ₀ = 1, and q ₀ = 1. Further, k = 1.

In the input layer, as an input value x _i , a value indicating fluctuations in the intake air temperature output from the fluctuation acquisition unit 13 (current value of the intake air temperature (x ₁ ), first-order differential value (x ₂ ), and A second-order differential value (x ₃ )) is input. These values are a kind of state quantities that affect the heat exchange by the heat exchange system 100 (for example, various state quantities that affect the cooling characteristics of the heat exchange system 100), but in addition to at least some of these values. Alternatively, another state quantity (a state quantity that affects heat exchange by the heat exchange system 100) may be input to the input layer. The state quantity includes a physical quantity, an operation quantity, a set value, and the like, and may be a current value, a differential value based on time, or the like. For example, various physical quantities of the refrigerant such as the condensation pressure detected by the pressure sensor SP1, the intake air temperature detected by the temperature sensor ST1, and the degree of superheat calculated above may be input as the state quantity. Furthermore, various set values such as the blown air temperature set value, the superheat degree set value, and the condensation pressure set value may be input as the state quantity. Further, as the state quantity, various manipulated variables (specifically, the manipulated variable of the past three manipulated variables input to the electronic expansion valve 106 and the past three manipulated variables inputted to the inverter 102). A value indicating a history, a change amount of the operation amount (current operation amount−the previous operation amount), a first-order differential value (time differentiation) of the time change of the operation amount of the past multiple times, a second-order differential value (time differentiation) ), Etc.) may be input. The operation amount input to the inverter 102 is the sum of the operation amount from the PID control unit 11 and the operation amount from the neural network 15 (details will be described later), but the operation amount input to the neural network 15 (inverter The operation amount to 102) may be one of the operation amounts in addition to the sum. In addition, the input layer has a current value (x ₁ ) of the intake air temperature, a first-order differential value (x ₂ ), and at least a first-order differential value (x ₂ ) among the second-order differential values (x ₃ ), Alternatively, a second-order differential value (x ₃ ) may be input.

The neural network 15 derives the output y ₁ based on the input value. For example, the output y ₁ is calculated by the above equations. The neural network 15 outputs the manipulated variable to the inverter 102 as the output y ₁ .

  Such a neural network 15 includes a deviation input to the PID control unit 11 (see the dotted arrow G1 in FIG. 2) and an operation amount output from the PID control unit 11 (see the dotted arrow G2 in FIG. 2). Any one of them is used as a teacher signal, and machine learning is performed so that the deviation or the operation amount is zero. For example, the neural network 15 performs machine learning so as to output an operation amount that makes the deviation or the operation amount zero. The machine learning is performed by, for example, an error back propagation method, and weight correction by the error back propagation method is performed by a gradient method. In this way, the result of machine learning is reflected in the neural network 15. The neural network 15 derives the operation amount based on the result of the machine learning. For example, the neural network 15 may stop learning or may always learn after learning the heat exchange system 100 to the user and learning to some extent.

  The learned neural network 15 outputs, to the inverter 102, an operation amount that cancels the influence (details will be described later) due to fluctuations in the intake air temperature.

The sum of the operation amount from the PID control unit 11 and the operation amount y ₁ from the neural network 15 is input to the inverter 102. In this way, the inverter 102 (in other words, the output of the compressor 101 (the flow rate of the refrigerant) and the blown air temperature) is controlled by the neural network 15 and the PID control unit 11.

  The neural network 15 may control the blower 107 (the amount of air to be blown) in accordance with control on the inverter 102. For example, when the flow rate of the refrigerant is increased by controlling the inverter 102 or the like (when the blown air is further cooled), the blower 107 may be controlled to increase the blown air in order to increase the cooling capacity. .

[Effects of this embodiment]
(1) The fluctuation of the intake air temperature adversely affects the heat exchange between the refrigerant and the air (control of the blown air temperature by the PID control unit 11). This is because the blown air temperature also depends on the intake air temperature. Note that the adverse effects described above are, for example, that the blown air temperature does not approach the target temperature, and the blown air temperature is not stable. In this embodiment, the neural network 15 learns a deviation input to the PID control unit 11 or an operation amount output from the PID control unit 11 as a teacher signal (learning to set the deviation or the operation amount to 0). The operation amount that cancels the influence on the heat exchange due to the fluctuation of the intake air temperature is derived and output. Therefore, adverse effects on heat exchange due to fluctuations in the intake air temperature can be reduced (ideally can be eliminated). The neural network 15 may perform machine learning to reduce the deviation or the operation amount, for example, and sequentially derive the operation amount so as to reduce the influence on the heat exchange due to the fluctuation of the intake air temperature. Output, and the adverse effects described above can be reduced.

  In the control by the controller 120, the transition of the blown air temperature by the control when the neural network 15 is present (control by the PID control unit 11 and the neural network 15) and the control when the neural network 15 is not present (control only by the PID control unit 11) The differences are shown in FIGS. 4B and 4C. Here, the fluctuation | variation of the suction air temperature shown to FIG. 4 (A) is given. The blown air temperature set value is 20 ° C. The neural network 15 here has been learned to some extent. As shown in FIG. 4, when there is a change in the intake air temperature, the blown air temperature is unstable due to the influence of the change only in the PID control (FIG. 4B). On the other hand, when the neural network 15 is controlled, the temperature of the blown air is stably changed (FIG. 4C). As described above, the control by the neural network 15 negates the adverse effect on the control of the blown air temperature due to the fluctuation of the intake air temperature, and can reduce the adverse effect on the heat exchange.

  Note that adverse effects on heat exchange due to fluctuations in the intake air temperature may be reduced by, for example, feedforward control, but it is difficult to set parameters for feedforward control (for example, no error is allowed). According to the neural network 15, it is possible to automatically learn the adverse effect on heat exchange due to fluctuations in the intake air temperature by machine learning, eliminating the need for know-how necessary for setting parameters, and solving the above disadvantages. Furthermore, although feedforward control cannot respond to the aging of the cooling characteristics of the heat exchange system 100, the neural network 15 learns sequentially (learning even after the heat exchange system 100 is installed and operated). It is possible to cope with the secular change. Furthermore, according to the neural network 15, the adverse effect on the heat exchange can be reduced compared to the feedforward control.

(2) Since the first-order differential value and the second-order differential value of the time variation of the fluctuation are input to the neural network 15 as values representing the fluctuation of the intake air temperature, the neural network 15 has the fluctuation of the intake air temperature. Can be captured effectively, and effective learning can be performed. Only one of the first-order differential value and the second-order differential value may be input to the neural network 15.

[Modifications, etc.]
The present invention is not limited to the above embodiment, and the above embodiment can be modified as appropriate. Although the example is shown below as a modification, the following modification can also combine at least one part. In addition, the configuration disclosed in this specification can be omitted as long as there is no technical problem.

(Modification 1)
The blown air temperature (set value) input and set from the operation unit 112 may be a room temperature, an intake air temperature, or the like. When setting the room temperature, a temperature sensor for detecting the room temperature may be provided, and the value of the room temperature detected by the temperature sensor may be used as a feedback value for PID control or the like. When the intake air temperature (the temperature is also close to room temperature) is set, the value of the intake air temperature detected by the temperature sensor ST1 may be used as a feedback value for PID control or the like.

(Modification 2)
Assuming that the evaporation pressure exceeds the upper limit pressure of the refrigerant that can be input to the compressor 101, a refrigerant bypass path (part of the refrigerant discharged from the compressor 101 is input to the compressor 101 again. Route) may be provided. In this case, an evaporation pressure adjusting valve (a valve for adjusting the flow rate of the refrigerant evaporated by the evaporator 108) is provided between the evaporator 108 and the compressor 101 on the circulation path L1, and at the time of bypass, the evaporation pressure adjusting valve The blown air temperature may be controlled by the opening degree. At this time, the evaporating pressure adjusting valve may be controlled by the neural network 15 and the PID control unit for the evaporating pressure adjusting valve. At this time, the neural network 15 performs learning or the like in the same manner as in the above embodiment so as to reduce the adverse effect on the control of the blown air temperature. Note that a neural network for the evaporation pressure regulating valve (learned so as to reduce the adverse effect similar to the neural network 15) and the neural network 15 for the inverter 102 may be prepared separately.

(Modification 3)
The neural network 15 controls at least one of the cooling water flow rate adjustment valve 105 and the blower 107 in place of or in addition to at least one of the inverter 102 and the electronic expansion valve 106. Also good. For example, the neural network 15 performs machine learning using a manipulated variable supplied to the cooling water flow rate adjustment valve 105 from a PID control unit that performs PID control of the cooling water flow rate adjustment valve 105 or a deviation input to the PID control unit as a teacher signal. A numerical value (such as a first-order differential value) representing the intake air temperature is input, and an operation amount is output to the cooling water flow rate adjustment valve 105.

(Modification 4)
Instead of the neural network 15, another control unit that performs machine learning may be provided.

(Modification 5)
The heat exchange system 100 may perform heating (in this case, the refrigerant circulation direction is reversed). The heat exchange system 100 only needs to use a heat exchangeable heat medium. In addition, the object to be cooled or heated by the heat exchange system 100 is not limited to air, but may be other gas, liquid, or the like. For example, the heat exchange system 100 may be used for a freezer. The heat exchange system is not limited to a direct expansion type cooling system, but may be another heat exchange system. In this case, the control objects by the neural network 15 and the PID control unit 11 are also changed as appropriate. However, it is desirable that the control target is an adjusting device (for example, an inverter that controls the compressor, various valves provided in the flow path of the heat medium) that adjusts the flow rate of the refrigerant used in the heat exchange system. Further, the object to be controlled is heat exchange (for example, control of the temperature of the fluid after the heat exchange by the PID control unit 11, the temperature of the fluid after the heat exchange and its fluctuation, or the degree of change in the temperature of the fluid changed by the heat exchange. It is desirable that the adjusting device (e.g., an inverter or an electronic expansion valve) that affects (controls heat exchange). The value input to the neural network 15 is a change in physical quantity that affects heat exchange (for example, temperature fluctuation of refrigerant flowing into the evaporator 108, temperature fluctuation of superheat degree, pressure at a predetermined position of refrigerant, etc.) ) (Current value, differential value, etc.) may be used.

(Modification 6)
Various values used in the controller 120 (for example, various values input to the PID control unit 11, the neural network 15, etc.) may be appropriately normalized (however, even if normalized, the values are converted into the values). There is no change).

(Modification 7)
The PID control is feedback control that feeds back a predetermined physical quantity (particularly, a physical quantity that varies according to the flow rate of the refrigerant adjusted by a device controlled by the PID control), and brings the physical quantity close to a target value. The PID control may be changed to another feedback control. Other feedback control includes PI (Proportional-Integral) control and the like.

(Modification 8)
The physical quantity (temperature or pressure) detected by each sensor may be calculated. For example, the refrigerant vapor temperature may be calculated based on the evaporation pressure or the like.

L1 ... circulation path, 11 ... PID control unit, 13 ... variation acquisition unit, 15 ... neural network, 100 ... heat exchange system, 101 ... compressor, 102 ... inverter , 103 ... Condenser, 104 ... Cooling water channel, 105 ... Cooling water flow rate adjustment valve, 106 ... Electronic expansion valve, 107 ... Blower, 108 ... Evaporator, 112 ... Operation unit, 120 ... controller, SP1 to SP2 ... pressure sensor, ST1 to ST3 ... temperature sensor

Claims (5)

A heat exchange system for performing heat exchange using a heat medium,
An adjusting device for controlling the heat exchange by adjusting the flow rate of the heat medium;
A feedback control unit that performs feedback control to output an operation amount to the adjustment device so that a deviation between the first physical quantity that changes according to the flow rate and a target value becomes 0;
An input value including a value representing a variation of the second physical quantity that affects the heat exchange is input, and is a machine learning unit that controls the adjusting device based on the input value, and the teacher learns the deviation or the operation amount. A machine learning unit that performs learning to reduce the deviation or the operation amount as a signal;
A heat exchange system comprising. The input value includes a first-order differential value and a second-order differential value of the time change of the second physical quantity,
The heat exchange system according to claim 1. The fluctuation of the second physical quantity is a fluctuation of the temperature of the fluid before the heat exchange.
The heat exchange system according to claim 1 or 2. A controller for controlling an adjustment device that controls the heat exchange by adjusting a flow rate of a heat medium used for heat exchange,
A feedback control unit that performs feedback control to output an operation amount to the adjustment device so that a deviation between the first physical quantity that changes according to the flow rate and a target value becomes 0;
An input value including a value representing a variation of the second physical quantity that affects the heat exchange is input, and is a machine learning unit that controls the adjusting device based on the input value, and the teacher learns the deviation or the operation amount. A machine learning unit that performs learning to reduce the deviation or the operation amount as a signal;
Controller with. A method for constructing a neural network for controlling an adjusting device for controlling the heat exchange by adjusting a flow rate of a heat medium used for heat exchange,
The deviation or the operation amount in the feedback control for outputting the operation amount to the adjusting device such that the deviation between the first physical quantity that changes according to the flow rate and the target value becomes 0 is used as a teacher signal. Having the neural network perform learning to reduce
The neural network receives an input value including a value representing a variation in the second physical quantity that affects the heat exchange, and controls the adjusting device based on the input value.
How to build a neural network.