Incubator with multiple independent temperature control box bodies and artificial intelligence temperature control system

Abstract

The invention discloses an incubator with multiple independent temperature control box bodies and an artificial intelligent temperature control system, and belongs to the technical field of automatic control. The artificial intelligence temperature control system comprises N artificial intelligence subsystems for respectively controlling the temperatures of the N box bodies, in each artificial intelligence subsystem, a subtracter is used for subtracting the set temperature in the box body i from the actually measured temperature to obtain an error signal, and a reinforcement learning module i generates the state of the box body i according to the error signal, a control strategy of a PID controller i for controlling an electronic expansion valve i or a heating wire i at the time t is generated according to the state, and the reinforcement learning module also generates a control strategy of an electronic change-over switch i at the time t according to the error signal. The temperature of the multiple box bodies can be controlled through one compressor, so that the cost is reduced, and the efficiency is improved.

Description

Technical Field

[0001] The present invention relates to an incubator with multiple independent temperature control boxes and an artificial intelligence temperature control system, belonging to the technical field of automatic control. Background Art

[0002] Traditional split-system systems rely on a single compressor for centralized cooling, unable to independently adjust cooling capacity for each zone, resulting in large temperature fluctuations. Dual-unit cascade compressors are costly and complex to maintain, with low energy efficiency in low-temperature conditions. Refrigerant flow control lacks precision, and the thermal expansion valve has a slow response speed, making it unable to adapt to dynamic load changes. Summary of the Invention

[0003] To overcome the shortcomings of the prior art, the present invention provides an incubator with multiple independently temperature-controlled boxes and an artificial intelligence temperature control system, which can control the temperatures of multiple boxes through a single compressor, thereby reducing costs and improving efficiency.

[0004] To achieve the above-mentioned object of the invention, the present invention provides an incubator with multiple independently temperature-controlled boxes, which includes: a refrigerant supply module, N boxes and an artificial intelligence temperature control system, wherein the i-th box is provided with a temperature sensor i, an evaporator i, a heating wire i and an electronic expansion valve i, and the refrigerant supply module is connected to the evaporator i through the electronic expansion valve i; the heating wire i is used to provide heat energy to the box i, and the temperature sensor i is configured to obtain a sequence temperature in the box i. The artificial intelligence temperature control system includes N artificial intelligence subsystems for respectively controlling the temperatures of the N boxes, and the i-th artificial intelligence subsystem includes a reinforcement learning module i, a PID controller i, an electronic conversion switch i and a subtractor i, and the subtractor i is used to subtract the set temperature in the box i from the measured temperature to obtain an error signal , reinforcement learning module i according to the error signal Generate the state of box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t ,in, is the control parameter vector of PID controller i at time t, They are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t; reinforcement learning module i also calculates the error signal Generate the control strategy of electronic switch i at time t ,i=1,…,N, Represents the control signal of electronic switch i at time t.

[0005] To achieve the above-mentioned purpose, the present invention also provides an artificial intelligence temperature control system, which includes N artificial intelligence subsystems for controlling the temperature of N boxes respectively. The i-th artificial intelligence subsystem includes a reinforcement learning module i, a PID controller i, an electronic switch i and a subtractor i. The subtractor i is used to subtract the set temperature in the box i from the measured temperature to obtain an error signal. , reinforcement learning module i according to the error signal Generate the state of box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t ,in, is the control parameter vector of PID controller i at time t, They are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t; reinforcement learning module i also calculates the error signal Generate the control strategy of electronic switch i at time t ,i=1,…,N, Represents the control signal of electronic switch i at time t.

[0006] Compared with the prior art, the incubator and artificial intelligence temperature control system with multiple independently temperature-controlled boxes provided by the present invention have a reinforcement learning module of the artificial intelligence subsystem of each box, which generates a control strategy for a PID controller for controlling an electronic expansion valve or a heating wire according to the state generated by the error signal; and generates a control strategy for an electronic switching switch according to the error signal to select the electronic expansion valve or the heating wire to be connected to the PID controller, thereby achieving the goal of controlling the temperature of multiple boxes through one compressor, thereby reducing costs and improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 Schematic diagram of the incubator with multiple independent temperature-controlled boxes provided by the first embodiment of the present invention.

[0008] Figure 2 This is a block diagram of the composition of the artificial intelligence subsystem provided by the first embodiment of the present invention.

[0009] Figure 3 Schematic diagram of a neural network of a reinforcement learning module constructed according to the first embodiment of the present invention. DETAILED DESCRIPTION

[0010] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0011] First embodiment

[0012] Figure 1 Schematic diagram of the incubator with multiple independent temperature control boxes provided by the first embodiment of the present invention. Figure 1 As shown, the incubator provided by the present invention with multiple independently temperature-controlled chambers includes a housing containing N independent chambers. Each chamber has an evaporator 6 disposed on the outside of one side wall and a heating wire (not shown) disposed on the outside of the other side wall or the outside of the bottom wall. The incubator is also equipped with N temperature sensors, each with a probe extending into a chamber. The incubator also includes a refrigerant supply module and N electronic expansion valves, such as electronic expansion valve 7, electronic expansion valve 8, and electronic expansion valve 9. The refrigerant supply module comprises a fixed-frequency compressor 1, a condenser 2, and a capillary tube 5. The fixed-frequency compressor 1 draws in low-temperature, low-pressure refrigerant gas and compresses it into high-temperature, high-pressure gaseous refrigerant. The condenser receives the high-temperature, high-pressure gaseous refrigerant and, through a copper tube fin structure and a fan, forces heat dissipation, condensing the refrigerant into a high-pressure liquid and releasing heat to the outside. The capillary tube 5 throttles and reduces the pressure on the high-pressure liquid refrigerant, converting it into low-temperature, low-pressure refrigerant. This refrigerant is then delivered to N evaporators through N electronic expansion valves. Within the evaporators, the low-temperature, low-pressure refrigerant absorbs heat from the corresponding chamber and evaporates into a gaseous state, simultaneously lowering the air temperature and condensing and dehumidifying the air. The evaporators are designed in a coiled configuration. Optionally, a drying tube 4 containing a desiccant (such as molecular sieve or silica gel) is provided between the condenser and the capillary tube to absorb moisture from the refrigerant. Moisture can freeze in low-temperature environments, blocking the capillary tube and electronic expansion valve, causing system failure. Furthermore, the acid generated by the reaction of moisture with the refrigerant can corrode the piping. The drying tube is equipped with a filter to intercept impurities such as metal debris, installation residue, and compressor wear dust. If impurities enter the capillary tube or fixed-frequency compressor, they can cause blockage or mechanical damage. The drying tube's filtering function prevents such problems. The drying tube also has liquid storage capacity. The evaporator is connected to the electronic expansion valve via a high-voltage pipe and to the fixed-frequency compressor via a low-pressure gas pipe 3.

[0013] Each box is further provided with a fan, such as fan 10, which is used to adjust the airflow in the box and further adjust the temperature at various locations in the box.

[0014] In the first embodiment, the incubator also includes an artificial intelligence temperature control system, which includes N artificial intelligence subsystems that control the temperature of N boxes respectively. Figure 2 Provide a detailed description of the composition and working process of each artificial intelligence subsystem.

[0015] Figure 2 This is a block diagram of the composition of the artificial intelligence subsystem provided by the first embodiment of the present invention. Figure 2 As shown, the i-th artificial intelligence subsystem includes a reinforcement learning module i, a PID controller i, an electronic switch i and a subtractor i. The subtractor i is used to subtract the set temperature in the box i from the measured temperature to obtain an error signal , reinforcement learning module i according to the error signal Generate the state of box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t ,in, is the control parameter vector of PID controller i at time t, They are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t; reinforcement learning module i also calculates the error signal Generate the control strategy of electronic switch i at time t ,i=1,…,N, Represents the control signal of electronic switch i at time t.

[0016] In the first embodiment, the electronic expansion valve i and the heating wire i are connected to the output end via an adder i. The adder i may also be replaced by a connection node.

[0017] In the first embodiment, , Where, They are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t; the electronic expansion valve or heating wire is controlled according to the control parameter vector In the first embodiment, the electronic expansion valve and the heating wire do not work at the same time. When the temperature needs to be lowered, the electronic expansion valve works, and when the temperature needs to be increased, the heating wire works.

[0018] Still Figure 2 As shown, the output of PID controller i at time t is ; Set temperature inside chamber i at time t and the measured temperature at time t The error is : , wherein, , ; ; is the error of the set temperature of the box i at time and the measured temperature at time ; is the error of the set temperature of the box i at time and the measured temperature at time ; is the error of the set temperature of the box i at time and the measured temperature at time , wherein y id is the set temperature within the set time length.

[0019] In matrix form, it is: , wherein, , , wherein T represents transposition.

[0020] Figure 3 is a schematic diagram of the neural network of the reinforcement learning module constructed in the first embodiment of the present application, as shown in Figure 3 , the neural network includes an input layer, a hidden layer and an output layer, the input layer includes 3 neurons, respectively inputting three elements in the vector ; the hidden layer includes 3 neurons; the output layer includes 4 neurons, wherein the first-3 neurons respectively output the proportional, integral and differential coefficients of the PID controller, and the fourth neuron outputs the state value function.

[0021] In the first embodiment, the reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i. It includes: S1-1: calculate the coefficient at time t according to the following formula : , wherein, is the output function of the jth neuron of the hidden layer at time t, and is the center and width of the jth Gaussian function of the neural network of the reinforcement learning module i at time t, the setting of the center determines how quickly the neural network adapts to the change of , which directly affects the speed of .​​​ is the weight between the jth neuron of the hidden layer of the neural network of the reinforcement learning module i and the nth neuron of the output layer at time t, which is used to adjust the connection strength between neurons to affect the speed of information flow, while the weight affects the computing resources, directly affecting the control consumption; in addition, the contribution of the weight to the total output signal affects the signal level; j = 1, 2, 3; n = 1, 2, 3, denotes the two-norm.

[0022] The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i comprises: S1-2: Calculate the state value function at time t according to the following formula : , wherein, is the weight between the jth neuron of the hidden layer of the neural network of the reinforcement learning module i and the Nth neuron of the output layer at time t, which is used to adjust the connection strength between neurons to affect the speed of information flow, while the weight affects the computing resources, directly affecting the control consumption; in addition, the contribution of the weight to the total output signal affects the signal level, N = 4.

[0023] The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i further comprises: S1-3: Construct the loss function according to the following formula : , wherein, is the set temperature of the box i at time t, is the measured temperature of the box i at time t.

[0024] The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i The reinforcement learning module i generates the control strategy of the PID controller i at time t according to the state of the box i further comprises: S1-4: Update the parameters according to the following formula: , , , , wherein, is the reward obtained by the PID controller i adopting ; and , 、 、 is the learning coefficient; is the time between the jth neuron in the hidden layer and the nth neuron in the output layer of the neural network of the reinforcement learning module i The weights when n=1, 2, 3; is the time between the jth neuron in the hidden layer and the Nth neuron in the output layer of the neural network of the reinforcement learning module i The weight when N=4; is the jth Gaussian function of the neural network of reinforcement learning module i at time The center of time; is the jth Gaussian function of the neural network of reinforcement learning module i at time bandwidth when Express The gradient, Express gradient; Express gradient; Express gradient; is the discount factor; is the time step.

[0025] In the first embodiment, ; ; ; ; .

[0026] Reinforcement learning module i according to the state of box i Generate the control strategy of PID controller i at time t Also includes: S1-5: Determine the loss function Is it the smallest? If not, , , , And returns to step S1-1; If, output 、 、 、 As the optimal parameter to calculate 、 、 .

[0027] In the first embodiment, the reinforcement learning module i also uses the error signal Generating control strategy of electronic switch i at time t Including: judging If it is negative, PID controller i is connected to electronic expansion valve i through electronic switch i, electronic expansion valve i works according to the control of PID controller i, and the low-temperature and low-pressure refrigerant provided by the refrigerant supply module through electronic expansion valve i is provided to the evaporator arranged outside the side wall of cabinet i, the low-temperature and low-pressure refrigerant absorbs the heat in cabinet i and evaporates into gas state in the evaporator, and the temperature in cabinet i is reduced; if it is positive, PID controller i is connected to electric heating wire i through electronic switch i, electric heating wire i works according to the control of PID controller i, and electric heating wire i heats cabinet i, and the temperature in cabinet i is increased; if , neither electronic expansion valve i nor electric heating wire i works.

[0028] Compared with the prior art, the split type independent temperature control device and the artificial intelligence temperature control system of the incubator provided by the first embodiment have the following beneficial effects: the state generated by the error signal is used to generate the control strategy of the PID controller of the electronic expansion valve of the cooling evaporator of the cabinet or the heating wire for heating the cabinet through the reinforcement learning module of the artificial intelligence control subsystem of each cabinet; the control strategy of the electronic switch is generated according to the error signal to select the electronic expansion valve or the heating wire connected to the PID controller, so that the temperature of multiple cabinets can be controlled by one compressor, thereby reducing the cost and improving the efficiency.

[0029] In addition, although the parameters of the neural network of the reinforcement learning module are dynamically updated over time using the gradient descent method, the initial selection of these parameters is crucial to achieving the desired results.

[0030] Second embodiment

[0031] The second embodiment of the present application only describes the different content from the first embodiment, and the same content is not repeated.

[0032] In the second embodiment, the reinforcement learning module i generates the control strategy of the PID controller i of the electronic expansion valve i or the heating wire i at time t according to the error signal The state of cabinet i According to the state Generating control strategy of electronic switch i at time t Including: S2-1: Calculate the state-action function at time t according to the following formula: , In the formula, PID controller i adopts control parameter vector , the state is transferred to state with a transition probability is the state value function at time t; denotes a discount factor; is the reward obtained by PID controller i with control parameter vector ; S2-2: improve according to the following formula: , where is the time step; S2-3: calculate the state value function at time according to the following formula: ; S2-4: determine whether is less than or equal to a threshold value , if so, output the control parameter vector of PID controller i at time ; if not, set , and then return to step S2-1.

[0033] Compared with the prior art, the artificial intelligence temperature control system provided by the second embodiment has the following beneficial effects: the reinforcement learning module of the artificial intelligence control subsystem of each cabinet generates a state according to an error signal, generates a control strategy of a PID controller of an electronic expansion valve of a cooling evaporator of the cabinet or a heating wire for warming up the cabinet according to the state, and generates a control strategy of an electronic switching switch according to the error signal to select the electronic expansion valve or the heating wire to be connected to the PID controller, so that the temperature of multiple cabinets can be controlled by one compressor, thereby reducing the cost and improving the efficiency.

[0034] The preferred embodiments of the disclosed application are only used to help explain the application, and the preferred embodiments do not describe all the details and do not limit the application to the specific embodiments. Obviously, according to the content of the specification, many modifications and changes can be made, and the specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their entire scope and equivalents.

Claims

1. An incubator with multiple independent temperature control boxes, characterized in that: include: A refrigerant supply module, N boxes, and an artificial intelligence temperature control system. The i-th box is provided with a temperature sensor i, an evaporator i, a heating wire i, and an electronic expansion valve i. The refrigerant supply module is connected to the evaporator i through the electronic expansion valve i. The heating wire i is used to provide heat energy to the box i. The temperature sensor i is configured to obtain the sequence temperature in the box i. The artificial intelligence temperature control system includes N artificial intelligence subsystems that control the temperature of the N boxes respectively. The i-th artificial intelligence subsystem includes a reinforcement learning module i, a PID controller i, an electronic conversion switch i, and a subtractor i. The subtractor i is used to subtract the set temperature in the box i from the measured temperature to obtain an error signal , reinforcement learning module i according to the error signal Generate the state of box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t ,in, is the control parameter vector of PID controller i at time t, are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t respectively; The reinforcement learning module i also uses the error signal Generate the control strategy of electronic switch i at time t , Represents the control signal of electronic switch i at time t, i=1,...,N.

2. The incubator with multiple independent temperature-controlled boxes according to claim 1, characterized in that: , Where, are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t respectively.

3. The incubator with multiple independent temperature-controlled boxes according to claim 2, characterized in that: ， Where, , and is the center and width of the jth Gaussian function of the neural network of reinforcement learning module i at time t; is the weight between the jth neuron in the hidden layer of the neural network of the reinforcement learning module and the nth neuron in the output layer at time t; j=1,2,3; n=1,2,3, represents the two-norm.

4. The incubator with multiple independent temperature-controlled boxes according to claim 3, characterized in that: Reinforcement learning module i calculates the state value function at time t according to the following formula : ， Where, is the weight between the jth neuron in the hidden layer and the Nth neuron in the output layer of the neural network of reinforcement learning module i at time t, where N=4.

5. The incubator with multiple independent temperature-controlled boxes according to claim 4, characterized in that: Reinforcement learning module i constructs the loss function according to the following formula : ， Where, is the set temperature inside the box i, is the measured temperature inside box i; Update the parameters according to the following formula: ， ， , ， Where, The control parameter vector is adopted for the PID controller i and the rewards received; 、 、 、 is the learning coefficient; is the time between the jth neuron in the hidden layer and the nth neuron in the output layer of the neural network of the reinforcement learning module i The weights when n=1, 2, 3; is the time between the jth neuron in the hidden layer and the Nth neuron in the output layer of the neural network of the reinforcement learning module i The weight when N=4; is the jth Gaussian function of the neural network of reinforcement learning module i at time The center of time; is the jth Gaussian function of the neural network of reinforcement learning module i at time bandwidth when Express The gradient, Express gradient; Express gradient; Express gradient; is the discount factor; is the time step.

6. The incubator with multiple independent temperature-controlled boxes according to claim 1, characterized in that: Reinforcement learning module i generates The state of the box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t include: S2-1: Calculate the state-action function at time t according to the following formula: ， Where, At time t, PID controller i adopts the control parameter vector , from the state Transfer to state The transition probability of is the state value function at time t; represents the discount factor; The control parameter vector is adopted for the PID controller i and the rewards received; S2-2: Improve according to the following formula : ， Where, is the time step; S2-3: Calculate the time according to the following formula The state value function at time: ； S2-4: Judgment Is it less than or equal to the threshold? , if so, then the output PID controller i at time The control parameter vector when If not, , Then return to step S2-1.

7. An artificial intelligence temperature control system, characterized in that: It includes N artificial intelligence subsystems that control the temperature of N boxes respectively. The i-th artificial intelligence subsystem includes a reinforcement learning module i, a PID controller i, an electronic switch i and a subtractor i. The subtractor i is used to subtract the set temperature in box i from the measured temperature to obtain an error signal , reinforcement learning module i according to the error signal Generate the state of box i , according to the status Generate the control strategy of PID controller i controlling electronic expansion valve i or heating wire i at time t ,in, is the control parameter vector of PID controller i at time t, They are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t; reinforcement learning module i also calculates the error signal Generate the control strategy of electronic switch i at time t ,i=1,…,N, Represents the control signal of electronic switch i at time t.

8. The artificial intelligence temperature control system according to claim 7, characterized in that: , Where, are the proportional coefficient, integral coefficient and differential coefficient of PID controller i at time t respectively.

9. The artificial intelligence temperature control system according to claim 8, characterized in that: ， Where, , and is the center and width of the jth Gaussian function of the neural network of reinforcement learning module i at time t; is the weight between the jth neuron in the hidden layer of the neural network of reinforcement learning module i and the nth neuron in the output layer at time t; j=1,2,3; n=1,2,3, represents the two-norm; Reinforcement learning module i calculates the state value function at time t according to the following formula : ， Where, is the weight between the jth neuron in the hidden layer and the Nth neuron in the output layer of the neural network of reinforcement learning module i at time t, where N=4.

10. The artificial intelligence temperature control system according to claim 9, characterized in that: Reinforcement learning module i constructs the loss function according to the following formula : ， Where, is the set temperature inside the box i, is the measured temperature inside box i; Update the parameters according to the following formula: ， ， ， Where, The control parameter vector is adopted for the PID controller i and the rewards received; 、 、 、 is the learning coefficient; is the time between the jth neuron in the hidden layer and the nth neuron in the output layer of the neural network of the reinforcement learning module i The weights when n=1, 2, 3; is the time between the jth neuron in the hidden layer and the Nth neuron in the output layer of the neural network of the reinforcement learning module i The weight when N=4; is the jth Gaussian function of the neural network of reinforcement learning module i at time The center of time; is the jth Gaussian function of the neural network of reinforcement learning module i at time bandwidth when Express The gradient, Express gradient; Express in pairs gradient; Express gradient; is the discount factor; is the time step.

Images