Cold therapy cabin refrigerating system energy-saving control algorithm randomly used by user
By optimizing the refrigeration system of the cryotherapy chamber using fuzzy control and neural network algorithms, the problem of frequent changes in heat load caused by random user use was solved, the refrigeration efficiency and system safety were improved, energy consumption was reduced, and the stability of cryotherapy quality and energy-saving effects were achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-02-29
- Publication Date
- 2026-04-17
AI Technical Summary
When the refrigeration system of the cryotherapy chamber is used intermittently by users, the heat load changes frequently, resulting in low refrigeration efficiency and high energy consumption. Existing technologies are unable to effectively control the superheat of the refrigerant at the evaporator outlet, which affects the quality of cryotherapy and the safety of the system.
A fuzzy control structure and a neural network-based fuzzy control algorithm are adopted. The superheat deviation and deviation change rate of the refrigerant at the evaporator outlet are used as inputs. The opening of the electronic expansion valve is adjusted by a fuzzy PID controller and a BP neural network to achieve real-time control of the superheat. A five-layer feedforward network model is constructed to optimize the operation of the refrigeration system.
It improves the safety and efficiency of the cryotherapy chamber's refrigeration system, reduces unnecessary energy consumption, adapts to changes in the cryotherapy chamber's load, and ensures stable and efficient system operation.
Smart Images
Figure CN121879087A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of refrigeration system control methods, specifically to an energy-saving control algorithm for a user-randomly used cryotherapy chamber refrigeration system. Background Technology
[0002] A stable and energy-efficient refrigeration system in the cryotherapy chamber is crucial for the quality of cryotherapy for patients and for achieving the goals of "carbon peaking" and "carbon neutrality" in meeting user energy needs. Patients entering the cryotherapy chamber experience random and intermittent use; the heat load of the refrigeration system varies widely and frequently. The superheat of the refrigerant at the evaporator outlet directly affects the cooling effect of the cryotherapy chamber. Based on these operational characteristics, to reduce unnecessary energy consumption, the operation of the refrigeration system is divided into an in-chamber idle phase and a patient treatment phase, with the superheat of each phase used as a control target. This aims to improve the safety and efficiency of the cryotherapy chamber's refrigeration system and reduce energy consumption. Summary of the Invention
[0003] The purpose of this invention is to provide an energy-saving control algorithm for the refrigeration system of a cryotherapy chamber used randomly by users, so as to solve the problems mentioned in the background art of improving the safety and refrigeration efficiency of the cryotherapy chamber refrigeration system and reducing energy consumption.
[0004] To achieve the above objectives, the present invention adopts the following technical solution:
[0005] An energy-saving control algorithm for a user-randomly used cryotherapy chamber refrigeration system includes a fuzzy control structure and a neural network-based fuzzy control structure. The control algorithm uses the refrigerant superheat deviation at the evaporator outlet and the rate of change of this deviation as input correlation quantities, wherein the opening degree of the electronic expansion valve is the output control quantity. A fuzzy PID controller is constructed through the above input and output quantities, and the membership center value, width, and control rules of the fuzzy PID control are adjusted online through a BP neural network.
[0006] The fuzzy control structure controls the dynamic characteristics of superheat in real time, effectively reducing the lag of sensors in the acquisition and transmission process. It adopts a dual-input single-output fuzzy control design, with the superheat deviation e of the refrigerant at the evaporator outlet and the rate of change of the deviation ec as input quantities, and the refrigerant temperature and pressure at the evaporator outlet as feedback quantities. After fuzzification, fuzzy inference and defuzzification in sequence, the opening degree u of the electronic expansion valve is obtained to complete the fuzzy control of superheat.
[0007] The fuzzification process maps inputs from the actual domain to linguistic variables in the fuzzy domain. The actual range of the input and output variables in the actual domain is [a, b], and their values are continuous. The fuzzy domain Y = [-N, -N+1...0..., N+1, N], and its values are discrete. The mapping is a linear mapping, and the linear mapping relationship is... x is a value in the fundamental universe of discourse, and y is a value in the fuzzy universe of discourse.
[0008] The fuzzy inference process involves fuzzifying the input, deriving the result through pre-defined fuzzy control rules, and finally outputting it to a fuzzy subset. Real-time input can then be queried. The conclusion of the fuzzy inference primarily depends on fuzzy implication relations. The fuzzy set composition operation rules and the fuzzy control rule statement IFAANDBTHENC determine the fuzzy implication relationship. The relationship is a ternary fuzzy implication, and the composition operation rule is not unique. Therefore, the Mamdani fuzzy inference method is adopted.
[0009] The defuzzification process relies on the fuzzy set output by fuzzy inference, which needs to be transformed into values within the actual universe of discourse before it can be applied to the electronic expansion valve controller. This process is called defuzzification, and the centroid method (weighted average method) is used to defuzzify the aforementioned output fuzzy quantities. Where u A (z i ) indicates that the output quantity is within the universe of discourse value z i It is the membership value of the corresponding membership function. After the value is processed by the centroid method and scaled by the proportional factor, it is applied to the electronic expansion valve. By adjusting its opening, the superheat of the evaporator outlet can be regulated and controlled.
[0010] The neural network-based fuzzy control structure is a fuzzy neural network based on the Mamdani structure. It uses a distributed neural network to represent each process of fuzzy control and consists of a five-layer feedforward network. The five-layer network model consists of an input layer, a fuzzification layer, a fuzzy rule layer, a normalization layer, and an output layer.
[0011] The input layer represents the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*, with 2 nodes; the fuzzification layer nodes represent the fuzzy linguistic variable values of the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*; (The last part, "NB," appears to be an unrelated fragment and is omitted from the translation.) e NM e NS e ZO e PS e PM e PB e NB ec NM ec NS ec ZOec PS ec PM ec PB ec Fourteen nodes; the fuzzification layer fuzzifies the values passed from the upper layer, calculates the overheating deviation, and the membership degree of the overheating deviation to the fuzzy sets of each linguistic variable. The membership function is Gaussian membership type. i = 1, 2, j = 1, 2, ..., 14; x i For the i-th input variable, c ij δ ij , respectively, are the center and width of the Gaussian membership function for the j-th linguistic variable.
[0012] The fuzzy rule layer assigns one adjustment rule to each node, resulting in 7*7=49 fuzzy control rules corresponding to 49 nodes. The fuzzy rule layer calculates the applicability of each rule. The normalization layer has the same number of nodes as the fuzzy rule layer, which is 49 nodes. (a) j Transform into a dimensionless expression The output layer has a system output quantity U and one node. The output layer performs defuzzification, which is a weighted sum of all rules. The weighting coefficients represent the applicability of the fuzzy rule normalization calculated by the normalization layer.
[0013] The parameters that need to be adjusted in the fuzzy neural network model include: the center value c of the Gaussian membership function of the fuzzification layer. ij Width δ ij There are a total of 28 layers, and the connection weights w between the normalization layer and the output layer are... ij A total of 49 parameters, 77 of which were determined through model training, were used to apply the debugged fuzzy controller to the refrigeration system of the cryotherapy chamber and control the superheat. The operating parameters of the refrigeration system were collected and the collected data were analyzed. The superheat deviation of the refrigerant at the evaporator outlet, e, the rate of change of the deviation, ec, and the opening degree of the electronic expansion valve, u, were used as training samples.
[0014] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0015] The energy-saving control algorithm of this invention uses the refrigerant superheat deviation at the evaporator outlet and the rate of change of this deviation as input correlation quantities, with the opening degree of the electronic expansion valve as the output control quantity. A fuzzy PID controller is constructed using the above input and output quantities, and the membership center value, width, and control rules of the fuzzy PID control are adjusted online using a BP neural network. This control enables the refrigeration system to track the superheat setpoint more quickly and stably, adapting to the load variation characteristics of the cryotherapy chamber. Based on the operating characteristics of the cryotherapy chamber, unnecessary energy consumption is reduced. The superheat at each operating stage of the cryotherapy chamber is used as the control target to improve the safety and refrigeration efficiency of the cryotherapy chamber refrigeration system and reduce energy consumption. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of the superheat fuzzy control of the present invention.
[0017] Figure 2 This is a schematic diagram of fuzzy reasoning in this invention.
[0018] Figure 3 This is a schematic diagram showing the location layout of the electronic expansion valve in the refrigeration system of the present invention.
[0019] Figure 4 This is a schematic diagram of the electronic expansion valve structure of the present invention;
[0020] Figure 5 This is a schematic diagram illustrating the operational characteristics of the cryotherapy chamber of the present invention. Detailed Implementation
[0021] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.
[0022] An energy-saving control algorithm for a user-randomly used cryotherapy chamber refrigeration system includes a fuzzy control structure and a neural network-based fuzzy control structure. The control algorithm uses the refrigerant superheat deviation at the evaporator outlet and the rate of change of this deviation as input correlation quantities, wherein the opening degree of the electronic expansion valve is the output control quantity. A fuzzy PID controller is constructed through the above input and output quantities, and the membership center value, width, and control rules of the fuzzy PID control are adjusted online through a BP neural network.
[0023] The fuzzy control structure controls the dynamic characteristics of superheat in real time, effectively reducing the lag of sensors in the acquisition and transmission process. It adopts a dual-input single-output fuzzy control design, with the superheat deviation e of the refrigerant at the evaporator outlet and the rate of change of the deviation ec as input quantities, and the refrigerant temperature and pressure at the evaporator outlet as feedback quantities. After fuzzification, fuzzy inference and defuzzification in sequence, the opening degree u of the electronic expansion valve is obtained to complete the fuzzy control of superheat.
[0024] The fuzzification process maps inputs from the actual domain to linguistic variables in the fuzzy domain. The actual range of the input and output variables in the actual domain is [a, b], and their values are continuous. The fuzzy domain Y = [-N, -N+1...0..., N+1, N], and its values are discrete. The mapping is a linear mapping, and the linear mapping relationship is... x is a value in the fundamental universe of discourse, and y is a value in the fuzzy universe of discourse.
[0025] The deviation of refrigerant superheat at the evaporator outlet, e, and the rate of change of this deviation, ec, have actual domains of discourse of [-6,6] and [-3,3], respectively.
[0026] The opening degree u of the electronic expansion valve is used as the output quantity, and its actual domain of discourse is [0,100]. Due to excessive changes in superheat, the valve opening degree of the electronic expansion valve will change excessively, resulting in violent superheat oscillations, which in turn will further cause the electronic expansion valve opening to become too large. In order to ensure the smoothness of valve action changes and the accuracy of superheat control, the fuzzy linguistic variables are divided into 7 levels, and a triangle is used as the membership function.
[0027] Then the linguistic scale and membership function of the refrigerant superheat deviation e at the evaporator outlet in the fuzzy domain are:
[0028]
[0029] The deviation 'e' is easily affected by many nonlinear factors near zero. Therefore, the slope of the membership function can be set larger to improve sensitivity and timely response. Thus, the subset width is reduced at zero.
[0030] The linguistic scale and membership function of the rate of change of refrigerant superheat deviation at the evaporator outlet in the fuzzy domain are as follows:
[0031]
[0032] Therefore, the output fuzzy linguistic variables are also divided into 7 levels: {Negative Large NB, Negative Medium NM, Negative Small NS, Zero ZR, Positive Small PS, Positive Medium PM, Positive Large PB}.
[0033] Fuzzy control rules
[0034] The quality of fuzzy control rule design directly affects the control quality of overheating. When the deviation is large, the control method should be selected with the goal of eliminating the deviation as quickly as possible; while when the error is small, the control method should be selected with the goal of system stability to prevent overshoot.
[0035] Its control principles are:
[0036] a. When the superheat error is very large or significant, the main approach is to increase the opening of the electronic expansion valve in order to increase the refrigerant flow and quickly reduce the superheat error.
[0037] b. When the superheat error is very small or minimal in the positive direction, while increasing the opening of the electronic expansion valve, it is also necessary to avoid system overshoot.
[0038] c. When the superheat error is very large or negative, the main approach is to reduce the opening of the electronic expansion valve in order to reduce the refrigerant flow and quickly reduce the superheat error.
[0039] e. When the superheat error is very small or negligible, reduce the opening of the electronic expansion valve while avoiding system overshoot.
[0040] Since the refrigerant superheat deviation e at the evaporator outlet and the rate of change ec of this deviation each have 7 states, a total of 7*7=49 fuzzy control rules need to be established. Therefore, the fuzzy control table is as follows:
[0041]
[0042] Principle of Evaporator Outlet Superheat Adjustment
[0043] The electronic expansion valve in the refrigeration system of the cryotherapy chamber is positioned and functions to monitor the refrigerant superheat at the evaporator outlet through pressure and temperature sensors when the cooling load inside the chamber changes. This allows for adjustment of the electronic expansion valve opening to regulate the refrigerant flow rate, preventing low evaporation heat exchange efficiency and the entry of the gas-liquid mixture into the compressor suction channel, thus ensuring stable and efficient operation of the cryotherapy chamber's refrigeration system.
[0044] The electronic expansion valves are respectively arranged on the evaporator EHX outlet channel and the pressure vessel PV outlet channel in the refrigeration system; EXV-1, EXV-2, and EXV-3 are arranged on the evaporator EHX outlet channel, and EXV-4 is arranged on the pressure vessel PV outlet channel. (See below) Figure 2 The bolded black section shows the location layout of the electronic expansion valve in the refrigeration system.
[0045] The evaporator EHX outlet superheat is a crucial operating parameter of the refrigeration system, directly affecting the refrigeration efficiency within the cryotherapy chamber and being related to the safe operation of the refrigeration system. The superheat is the difference between the actual temperature of the refrigerant at the evaporator outlet and its saturation temperature under current operating conditions. Insufficient superheat at the evaporator outlet will cause liquid refrigerant to enter the compressor return gas passage, resulting in "liquid slugging"; excessive superheat will increase energy consumption.
[0046] The valve structure of the electronic expansion valve shown is as follows: Figure 4 Simplified diagram of electronic expansion valve structure:
[0047]
[0048] A = π × D 2 / 4 H≥H max
[0049] A represents the throttling area at the current opening of the throttle valve, and H represents the distance between the valve needle and the valve orifice. D is the valve needle cone angle, and D is the valve needle orifice diameter.
[0050] The refrigerant pressure changes before and after passing through the electronic expansion valve:
[0051] ΔP=P c -P e
[0052] ΔP is the pressure drop across the electronic expansion valve, P c P is the condensation pressure. e This is the evaporation pressure.
[0053] Then, at the current opening degree, the refrigerant flow rate q through the electronic expansion valve is... s :
[0054]
[0055] F is the refrigerant trimming factor, K is the refrigerant constant, and C is the refrigerant constant. e Here, A is the flow coefficient, and A is the throttling area of the throttle valve at its current opening.
[0056] Heat exchange of refrigerant within the evaporator EHX:
[0057] Q e =q s ×(h e,i -h e,o )×F=k×A evap ×(T w -T m )
[0058]
[0059] h e,iThe enthalpy value at the evaporator inlet (kJ / kg); h e,o Δ is the enthalpy at the evaporator outlet (kJ / kg); k is the heat transfer coefficient of the evaporator; A evap T represents the heat transfer area of the evaporator. w Evaporator tube inner wall temperature; T m T represents the average temperature of the refrigerant. e,i T is the temperature of the refrigerant at the evaporator inlet. e,o This refers to the temperature of the refrigerant at the evaporator outlet.
[0060] As can be seen from the above, the flow rate q of the refrigerant flowing through the evaporator can be controlled by adjusting the opening of the electronic expansion valve. s This, in turn, affects the refrigerant temperature T at the evaporator outlet. e,o This system aims to control the superheat at the evaporator outlet. Specifically, electronic expansion valves EXV-1, EXV-2, and EXV-3 are used to control the evaporator outlet superheat in the -40℃, -80℃, and -110℃ refrigeration circulation channels, respectively. Pressure sensor P_3 and temperature sensor T_3 collect the fluid state of the evaporator outlet return steam, participating in closed-loop control to provide a rapid response to evaporator outlet superheat and improve refrigeration system performance. Furthermore, when the refrigeration load in the cryotherapy chamber changes, the above adjustments adapt to the load changes, ensuring efficient and safe system operation.
[0061] There are generally two methods for fuzzy inference: 1) real-time calculation, which uses a high-performance computer to calculate in real time; 2) offline query, which uses defined fuzzy rules to pre-calculate the corresponding output for different input quantities and saves the data offline in a query table, and then uses method 1 for inference.
[0062] The fuzzy inference process involves fuzzifying the input, deriving the result through pre-defined fuzzy control rules, and finally outputting it to a fuzzy subset. In practice, real-time input can be queried, and the conclusion of the fuzzy inference mainly depends on the fuzzy implication relation. The fuzzy set composition operation rules and the fuzzy control rule statement IFAANDBTHENC determine the fuzzy implication relationship. The relationship is a ternary fuzzy implication, and the composition operation rule is not unique. Therefore, the Mamdani fuzzy inference method is adopted.
[0063] Mamdani's fuzzy inference method uses the Cartesian product of A and B to represent the fuzzy implication relation A→B, that is:
[0064] R = A → B = A × B
[0065] R(x,y)=A(x)∧B(y)
[0066] ∧; Cartesian product, take the minimum.
[0067] Fuzzy syllogisms are often referred to as hypothetical syllogisms that affirm the antecedent, i.e., given the implication relation A→B and the relation R(X,Y), for the known A... * A * ∈X, we can deduce the conclusion is B. * B * ∈Y, the expression is:
[0068] B * =A * ∩R(X,Y)
[0069] Where “∩” represents the composition calculation, that is:
[0070]
[0071] or:
[0072]
[0073] refers to fuzzy set A * The height of the intersection with A is represented as:
[0074] α=H(A*IA) can be regarded as A * The degree of matching with A, i.e., membership degree.
[0075] Based on the Mamdani fuzzy reasoning method described above, to find B * First, the matching degree α (maximum intersection value) should be calculated. Then, α should be used to select the split B, and the minimum value should be taken to obtain B. * Therefore, this method is also known as the top-cutting method.
[0076] The defuzzification process relies on the fuzzy set output by fuzzy inference, which needs to be transformed into values within the actual universe of discourse before it can be applied to the electronic expansion valve controller. This process is called defuzzification, and the centroid method (weighted average method) is used to defuzzify the aforementioned output fuzzy quantities. Where u A (z i ) indicates that the output quantity is within the universe of discourse value z i It is the membership value of the corresponding membership function. After the value is processed by the centroid method and scaled by the proportional factor, it is applied to the electronic expansion valve. By adjusting its opening, the superheat of the evaporator outlet can be regulated and controlled.
[0077] The neural network-based fuzzy control structure is a fuzzy neural network based on the Mamdani structure. It uses a distributed neural network to represent each process of fuzzy control and consists of a five-layer feedforward network. The five-layer network model consists of an input layer, a fuzzification layer, a fuzzy rule layer, a normalization layer, and an output layer.
[0078] The input layer represents the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*, with 2 nodes; the fuzzification layer nodes represent the fuzzy linguistic variable values of the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*; (The last part, "NB," appears to be an unrelated fragment and is omitted from the translation.) e NM e NS e ZO e PS e PM e PB e NB ec NM ec NS ec ZO ec PS ec PM ec PB ec Fourteen nodes; the fuzzification layer fuzzifies the values passed from the upper layer, calculates the overheating deviation, and the membership degree of the overheating deviation to the fuzzy sets of each linguistic variable. The membership function is Gaussian membership type. x i For the i-th input variable, c ij δ ij , respectively, are the center and width of the Gaussian membership function for the j-th linguistic variable.
[0079] The fuzzy rule layer assigns one adjustment rule to each node, resulting in 7*7=49 fuzzy control rules corresponding to 49 nodes. The fuzzy rule layer calculates the applicability of each rule. The normalization layer has the same number of nodes as the fuzzy rule layer, which is 49 nodes. (a) j Transform into a dimensionless expression The output layer has a system output quantity U and one node. The output layer performs defuzzification, which is a weighted sum of all rules. The weighting coefficients represent the applicability of the fuzzy rule normalization calculated by the normalization layer.
[0080] The parameters that need to be adjusted in the fuzzy neural network model include: the center value c of the Gaussian membership function of the fuzzification layer. ij Width δ ij There are a total of 28 layers, and the connection weights w between the normalization layer and the output layer are... ij A total of 49 parameters, 77 of which were determined through model training, were used to apply the debugged fuzzy controller to the refrigeration system of the cryotherapy chamber and control the superheat. The operating parameters of the refrigeration system were collected and the collected data were analyzed. The superheat deviation of the refrigerant at the evaporator outlet, e, the rate of change of the deviation, ec, and the opening degree of the electronic expansion valve, u, were used as training samples.
[0081] Features of cryotherapy chamber users
[0082] The cooling load Q0 of the cryotherapy chamber includes: cooling load due to chamber body loss Q1, heat leakage from opening the chamber door Q2, heat leakage from the chamber space Q3, radiative heat exchange between the chamber and the environment Q4, and the cooling load within the chamber Q5. The heat leakage from opening the chamber door Q2 and the cooling load within the chamber Q5 apply heat load to the refrigeration system during patient cryotherapy. After treatment, the heat load of the refrigeration system mainly consists of: cooling load due to chamber body loss Q1, heat leakage from the chamber space Q3, and radiative heat exchange between the chamber and the environment Q4. To save energy and reduce consumption, the cooling capacity of the refrigeration system should be controlled separately for the heat load during the unloaded phase and the patient treatment phase. That is, during the unloaded phase inside the chamber, the cooling capacity is matched with the chamber's lost cooling load Q1, the heat leakage of the chamber space Q3, and the radiative heat exchange of the chamber to the environment Q4, so as to control the temperature inside the chamber at the preset temperature T1; during the patient treatment phase, the cooling capacity is matched with the chamber's lost cooling load Q1, the heat leakage of the chamber door Q2, the heat leakage of the chamber space Q3, the radiative heat exchange of the chamber to the environment Q4, and the cold therapy load inside the cold therapy chamber Q5, so as to control the temperature inside the chamber at the treatment temperature T2.
[0083] When the cabin temperature is at the preset temperature T1, the refrigeration system operates intermittently, stabilizing the cabin temperature within a certain range of T1, where T1 is greater than the saturation temperature T of the refrigerant in the evaporator of the current refrigeration cycle. e In this state, by adjusting the superheat at the evaporator outlet according to the above-mentioned principle, the opening of the electronic expansion valve is adjusted so that Q in formula 4 is... e Match the cabin's cooling load Q1, the cabin's space heat leakage Q3, and the cabin's radiative heat exchange with the environment Q4.
[0084] When the treatment temperature T1 inside the chamber is reached, the refrigeration system operates continuously. The treatment temperature T1 inside the chamber is close to the saturation temperature inside the evaporator and is maintained within a certain range. T1 = saturation temperature T of the refrigerant in the evaporator in the current refrigeration cycle channel. e In this state, by adjusting the superheat at the evaporator outlet according to the above-mentioned principle, the opening of the electronic expansion valve is adjusted so that Q in formula 4 is... e The matching chamber's cooling load Q1, heat leakage when opening the chamber door Q2, heat leakage from the chamber space Q3, heat exchange from the chamber to the environment via radiation Q4, and the cooling load inside the cooling chamber Q5 are all considered.
[0085] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the present invention without departing from its novel spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.
Claims
1. An energy-saving control algorithm for a user-randomly used cryotherapy chamber refrigeration system, characterized in that, It includes fuzzy control structures and neural network-based fuzzy control structures; the control algorithm uses the refrigerant superheat deviation at the evaporator outlet and the rate of change of this deviation as input related quantities, and the opening degree of the electronic expansion valve as the output control quantity; a fuzzy PID controller is constructed through the above input and output quantities, and the membership center value, width and control rules of the fuzzy PID control are adjusted online through a BP neural network.
2. The energy-saving control algorithm for the cooling system of the user-randomly-used cryotherapy chamber according to claim 1, characterized in that, The fuzzy control structure controls the dynamic characteristics of superheat in real time, effectively reducing the lag of sensors in the acquisition and transmission process. It adopts a dual-input single-output fuzzy control design, with the superheat deviation e of the refrigerant at the evaporator outlet and the rate of change of the deviation ec as input quantities, and the refrigerant temperature and pressure at the evaporator outlet as feedback quantities. After fuzzification, fuzzy inference and defuzzification, the opening degree u of the electronic expansion valve is obtained to complete the fuzzy control of superheat.
3. The energy-saving control algorithm for the cooling system of the user-randomly-used cryotherapy chamber according to claim 2, characterized in that, The fuzzification process maps inputs from the actual domain to linguistic variables in the fuzzy domain. The actual range of the input and output variables in the actual domain is [a, b], and their values are continuous. The fuzzy domain Y = [-N, -N+1...0..., N+1, N], and its values are discrete. The mapping is a linear mapping, and the linear mapping relationship is... x is a value in the fundamental universe of discourse, and y is a value in the fuzzy universe of discourse.
4. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 2, characterized in that, The fuzzy inference process fuzzifies the input, derives the result through pre-defined fuzzy control rules, and finally outputs it to a fuzzy subset. Real-time input can then be queried. The conclusion of the fuzzy inference primarily depends on fuzzy implication relations. The rules for composition between fuzzy sets, and the fuzzy control rule statement IF A AND B THEN C, determine the fuzzy implication relationship. For ternary fuzzy implication relations, the Mamdani fuzzy inference method is adopted.
5. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 2, characterized in that, The defuzzification process relies on the fact that the fuzzy set output by fuzzy inference needs to be transformed into values within the actual universe of discourse before it can be applied to the electronic expansion valve controller. This process is called defuzzification, and the centroid method (weighted average method) is used to defuzzify the aforementioned output fuzzy quantities. Where u A (z i ) indicates that the output quantity is within the universe of discourse value z i It is the membership value of the corresponding membership function. After the value is processed by the centroid method and scaled by the proportional factor, it is applied to the electronic expansion valve. By adjusting its opening, the superheat of the evaporator outlet can be regulated and controlled.
6. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 1, characterized in that, The neural network-based fuzzy control structure is a fuzzy neural network based on the Mamdani structure, which uses a distributed neural network to represent each process of fuzzy control, and consists of a five-layer feedforward network. The 5-layer network model consists of: input layer, fuzzification layer, fuzzy rule layer, normalization layer, and output layer.
7. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 6, characterized in that, The input layer represents the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*, with 2 nodes; the fuzzification layer nodes represent the fuzzy linguistic variable values of the refrigerant superheat deviation *e* at the evaporator outlet and the rate of change of this deviation *ec*; (The last part, "NB," appears to be an unrelated fragment and is omitted from the translation.) e NM e NS e ZO e PS e PM e PB e NB ec NM ec NS ec ZO ec PS ec PM ec PB ec Fourteen nodes; the fuzzification layer fuzzifies the values passed from the upper layer, calculates the overheating deviation, and the membership degree of the overheating deviation to the fuzzy sets of each linguistic variable. The membership function is Gaussian membership type. x i For the i-th input variable, c ij δ ij , respectively, are the center and width of the Gaussian membership function for the j-th linguistic variable.
8. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 6, characterized in that, The fuzzy rule layer assigns one adjustment rule to each node, resulting in 7*7=49 fuzzy control rules corresponding to 49 nodes. The fuzzy rule layer calculates the applicability of each rule. The normalization layer has the same number of nodes as the fuzzy rule layer, which is 49 nodes. (a) j Transform into a dimensionless expression The output layer has a system output quantity U and one node. The output layer performs defuzzification, which is a weighted sum of all rules. The weighting coefficients represent the applicability of the fuzzy rule normalization calculated by the normalization layer.
9. The energy-saving control algorithm for the cooling system of the user-randomly used cryotherapy chamber according to claim 8, characterized in that, The parameters to be adjusted in the fuzzy neural network model are: the center value c of the Gaussian membership function in the fuzzy layer ij , the width δ ij , the connection weight w of the normalized layer and the output layer, a total of 49, and 77 parameters determined through model training. ij The debugged fuzzy controller is applied to the refrigeration system of the cold therapy cabin to control the superheat degree, the operating parameters of the refrigeration system are collected, the collected data are analyzed, the deviation e of the refrigerant superheat degree at the outlet of the evaporator and the change rate ec of the deviation and the opening degree u of the electronic expansion valve are used as training samples.