Electronic expansion valve control method based on deep dynamic fuzzy neural network

Abstract

The invention relates to the technical field of electronic expansion valve control, and particularly discloses an electronic expansion valve control method based on a deep dynamic fuzzy neural network, and the method comprises the steps: obtaining the operation data of a heat pump unit; normalizing the operation data of the heat pump unit to obtain normalized operation data of the heat pump unit, and dividing the normalized operation data of the heat pump unit into a training set and a test set; constructing a deep dynamic fuzzy neural network model, and optimizing various parameters of the deep dynamic fuzzy neural network model by using a genetic algorithm and the data in the training set to obtain a trained deep dynamic fuzzy neural network model; and inputting data in the test set into the trained deep dynamic fuzzy neural network model for prediction so as to output an optimal control strategy of the electronic expansion valve. The deep dynamic fuzzy neural network model is used for effectively controlling the electronic expansion valve, so that the overshoot of the heat pump system is small, the response speed is high, and the operation performance of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic expansion valve control, and more specifically, to an electronic expansion valve control method based on a deep dynamic fuzzy neural network. Background Art

[0002] The promotion and application of variable frequency compressors also promotes the expansion of the heat pump market. Electronic expansion valves are increasingly used as key throttling components of heat pump systems. The quality of their performance characteristics is related to the energy saving level of a heat pump system. For traditional thermal expansion valves, due to their own structural reasons, they have a long delay characteristic that affects the response performance of the system and is difficult to effectively adjust the flow in real time with the compressor displacement, which can easily lead to system instability and vibration and may even damage the compressor, thus restricting its large-scale and wide-ranging application.

[0003] The biggest feature of the electronic expansion valve is the timeliness of flow regulation. It can match the change of compressor displacement in time. Its control action rules are determined by the program and have great flexibility. In order to maximize the energy-saving effect of the heat pump, the present invention proposes an electronic expansion valve control method based on a deep dynamic fuzzy neural network, which optimizes the opening control method of the electronic expansion valve during operation, making the air-conditioning system more energy-efficient and greener. Summary of the invention

[0004] In view of the above-mentioned problems existing in the prior art, the present invention provides an electronic expansion valve control method based on a deep dynamic fuzzy neural network, which uses a deep dynamic fuzzy neural network model to effectively control the electronic expansion valve, thereby reducing the overshoot of the heat pump system, speeding up the response speed, and improving the system operation performance.

[0005] As a first aspect of the present invention, a method for controlling an electronic expansion valve based on a deep dynamic fuzzy neural network is provided. The method for controlling an electronic expansion valve based on a deep dynamic fuzzy neural network comprises the following steps:

[0006] Step S1: Acquire the operation data of the heat pump unit, wherein the operation data of the heat pump unit includes the operation frequency of the compressor, the opening degree of the electronic expansion valve, the evaporator outlet temperature and the evaporator inlet temperature;

[0007] Step S2: normalizing the heat pump unit operation data to obtain normalized heat pump unit operation data, and dividing the normalized heat pump unit operation data into a training set and a test set;

[0008] Step S3: constructing a deep dynamic fuzzy neural network model, and optimizing various parameters of the deep dynamic fuzzy neural network model using a genetic algorithm and the data in the training set to obtain a trained deep dynamic fuzzy neural network model;

[0009] Step S4: inputting the data in the test set into the trained deep dynamic fuzzy neural network model for prediction to output the optimal control strategy for the electronic expansion valve.

[0010] Furthermore, the formula for the normalization process is as follows:

[0011]

[0012] Among them, x i is the original heat pump unit operation data, x min is the minimum value in the original heat pump unit operation data, x max is the maximum value in the original heat pump unit operation data, It is the normalized operating data of the heat pump unit.

[0013] Furthermore, the step of dividing the normalized heat pump unit operation data into a training set and a test set further includes:

[0014] The normalized heat pump unit operation data is divided into a training set and a test set according to a ratio of 7:3.

[0015] Furthermore, the deep dynamic fuzzy neural network model includes an input layer, a membership function layer, a T-norm layer, a normalization layer and an output layer; wherein,

[0016] The input layer is the input port of data;

[0017] The formula of the membership function layer is:

[0018]

[0019] in, is the normalized heat pump unit operation data, i = {1, 2, ..., r} is the amount of input data, j = {1, 2, ..., u} represents the number of rules, u ij and c ij are the membership functions and their centers, σ ij is the width of the Gaussian function at different layers;

[0020] The T-norm layer, its j layer The formula is:

[0021]

[0022] The normalization layer, its j layer The formula is:

[0023]

[0024] The output layer has an output function y(x) as follows:

[0025]

[0026] Among them, w k is the weight of the kth layer, is the output function of the kth layer.

[0027] Furthermore, the process of the genetic algorithm includes:

[0028] S1.1: Encoding of membership function, using triangular membership function, with Z-type and S-type functions used at the beginning and end respectively, and using the intermediate parameters to be optimized as the triangle vertices to avoid gaps;

[0029] S1.2: Initialize the population, set the number of intersections, population size and number of iterations;

[0030] S1.3: Evaluate the fitness of individuals in the population. The fitness formula is:

[0031]

[0032] Among them, f I is the fitness function of the genetic algorithm, m i and l i is the value of the x-coordinate and the y-coordinate; the distance calculation formula between two adjacent nodes and three adjacent nodes is:

[0033]

[0034] Where: d 1i is the distance between any two adjacent nodes in the path, d 2i is the distance between any three adjacent nodes in the path, and n is the total number of nodes in the total path;

[0035] S1.4: Based on the population fitness calculated from the initial population, the optimal individual set and the fitness set of each generation are obtained;

[0036] S1.5: Select a two-point crossover operator to perform calculations to obtain a new generation of populations, and then mutate the new generation of populations to obtain a new generation of populations;

[0037] S1.6: To avoid the loss of the best individual, the calculation results are stored, including each initial population, each iteration population information, and each iteration population individual information, so as to obtain the optimal parameters.

[0038] Furthermore, the step of obtaining the heat pump unit operation data further includes:

[0039] The operation data of the heat pump unit is collected through the wire controller and various sensors.

[0040] The electronic expansion valve control method based on deep dynamic fuzzy neural network provided by the present invention has the following beneficial effects:

[0041] (1) Using dynamic structures to optimize the membership function layer on the traditional fuzzy neural network can enhance the model's ability to process data;

[0042] (2) Use genetic algorithms to optimize the parameters of fuzzy neural networks and improve the algorithm's operating speed;

[0043] (3) Using the optimized deep dynamic fuzzy neural network model to effectively control the electronic expansion valve can reduce the overshoot of the heat pump system, speed up the response speed, and improve the system operation performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation of the present invention.

[0045] Figure 1 A flow chart of the electronic expansion valve control method based on a deep dynamic fuzzy neural network provided by the present invention.

[0046] Figure 2 This is a structural diagram of the deep dynamic fuzzy neural network model provided by the present invention.

[0047] Figure 3 The present invention provides a flow chart of the genetic algorithm. DETAILED DESCRIPTION

[0048] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of an electronic expansion valve control method based on a deep dynamic fuzzy neural network proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0049] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so as to describe the embodiments of the present invention described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0050] In this embodiment, a method for controlling an electronic expansion valve based on a deep dynamic fuzzy neural network is provided. Figure 1 As shown, the electronic expansion valve control method based on deep dynamic fuzzy neural network includes the following steps:

[0051] Step S1: Acquire the operation data of the heat pump unit, wherein the operation data of the heat pump unit includes the operation frequency of the compressor, the opening degree of the electronic expansion valve, the evaporator outlet temperature and the evaporator inlet temperature;

[0052] It should be noted that the operation data of the heat pump unit is collected through a wire controller and various sensors.

[0053] Step S2: normalizing the heat pump unit operation data to obtain normalized heat pump unit operation data, and dividing the normalized heat pump unit operation data into a training set and a test set;

[0054] Preferably, to avoid the input data range x i The difference is large, which affects the final training effect. The formula for the normalization process is as follows:

[0055]

[0056] Among them, x i is the original heat pump unit operation data, x min is the minimum value in the original heat pump unit operation data, x max is the maximum value in the original heat pump unit operation data, It is the normalized operating data of the heat pump unit.

[0057] It should be noted that the normalized heat pump unit operating data needs to be Denormalize to get the actual data. The denormalization formula is:

[0058]

[0059] Among them, xi ' is the data after denormalization.

[0060] Preferably, the step of dividing the normalized heat pump unit operation data into a training set and a test set further includes:

[0061] The normalized heat pump unit operation data is divided into a training set and a test set according to a ratio of 7:3.

[0062] Step S3: constructing a deep dynamic fuzzy neural network model, and optimizing various parameters of the deep dynamic fuzzy neural network model using a genetic algorithm and the data in the training set to obtain a trained deep dynamic fuzzy neural network model;

[0063] Preferably, if Figure 2 As shown, the deep dynamic fuzzy neural network model includes an input layer, a membership function layer, a T-norm layer, a normalization layer and an output layer; wherein,

[0064] The input layer is the input port of data;

[0065] The formula of the membership function layer is:

[0066]

[0067] in, is the normalized heat pump unit operation data, i = {1, 2, ..., r} is the amount of input data, j = {1, 2, ..., u} represents the number of rules, u ij and c ij are the membership functions and their centers, σ ij is the width of the Gaussian function at different layers;

[0068] It should be noted that, in the membership function layer, different membership functions are set for input variables in each layer to avoid the situation where the signal widths are the same, so that the model has the ability to process wide-area data.

[0069] The T-norm layer, its j layer The formula is:

[0070]

[0071] The normalization layer, its j layer The formula is:

[0072]

[0073] The output layer has an output function y(x) as follows:

[0074]

[0075] Among them, w k is the weight of the kth layer, is the output function of the kth layer.

[0076] Preferably, if Figure 3 As shown, the process of the genetic algorithm includes:

[0077] S1.1: Encoding of membership function, using triangular membership function, with Z-type and S-type functions used at the beginning and end respectively, and using the intermediate parameters to be optimized as the triangle vertices to avoid gaps;

[0078] S1.2: Initialize the population, set the number of intersections, population size and number of iterations;

[0079] S1.3: Evaluate the fitness of individuals in the population. The fitness formula is:

[0080]

[0081] Among them, f I is the fitness function of the genetic algorithm, m i and l i is the value of the x-coordinate and the y-coordinate; the distance calculation formula between two adjacent nodes and three adjacent nodes is:

[0082]

[0083] Where: d 1i is the distance between any two adjacent nodes in the path, d 2i is the distance between any three adjacent nodes in the path, and n is the total number of nodes in the total path;

[0084] S1.4: Based on the population fitness calculated from the initial population, the optimal individual set and the fitness set of each generation are obtained;

[0085] S1.5: Select a two-point crossover operator to perform calculations to obtain a new generation of populations, and then mutate the new generation of populations to obtain a new generation of populations;

[0086] S1.6: To avoid the loss of the best individual, the calculation results are stored, including each initial population, each iteration population information, and each iteration population individual information, so as to obtain the optimal parameters.

[0087] Step S4: inputting the data in the test set into the trained deep dynamic fuzzy neural network model for prediction to output the optimal control strategy for the electronic expansion valve.

[0088] It should be noted that the core of the deep dynamic fuzzy neural network model is the membership function parameters of the fuzzy set, and these parameters determine the performance and effect of the model. The genetic algorithm mainly embodies the theory of "survival of the fittest", which integrates the basic rules of the biological evolution process into feasible solutions, and randomly exchanges information in the feasible solutions. The superior parts of the evolution are preserved and combined, and then better new individuals are continuously produced. The new individuals retain the excellent genes of the previous generation of individuals, and continuous updating and iteration allows the entire population to evolve.

[0089] The electronic expansion valve control method based on deep dynamic fuzzy neural network provided by the embodiment of the present invention first uses dynamic structure on the traditional fuzzy neural network to optimize the membership function layer and enhance the model's data processing capability. Then, the genetic algorithm is used to optimize the fuzzy neural network parameters to improve the algorithm's operating rate. Finally, the deep dynamic fuzzy neural network model is used to effectively control the electronic expansion valve, which can reduce the overshoot of the heat pump system, speed up the response speed, and improve the system's operating performance.

[0090] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. An electronic expansion valve control method based on deep dynamic fuzzy neural network, characterized in that: The electronic expansion valve control method based on deep dynamic fuzzy neural network includes the following steps: Step S1: Acquire the operation data of the heat pump unit, wherein the operation data of the heat pump unit includes the operation frequency of the compressor, the opening degree of the electronic expansion valve, the evaporator outlet temperature and the evaporator inlet temperature; Step S2: normalizing the heat pump unit operation data to obtain normalized heat pump unit operation data, and dividing the normalized heat pump unit operation data into a training set and a test set; Step S3: constructing a deep dynamic fuzzy neural network model, and optimizing various parameters of the deep dynamic fuzzy neural network model using a genetic algorithm and the data in the training set to obtain a trained deep dynamic fuzzy neural network model; Step S4: inputting the data in the test set into the trained deep dynamic fuzzy neural network model for prediction to output the optimal control strategy for the electronic expansion valve.

2. The electronic expansion valve control method based on deep dynamic fuzzy neural network according to claim 1 is characterized in that: The formula for the normalization process is as follows: Among them, x i is the original heat pump unit operation data, x min is the minimum value in the original heat pump unit operation data, x max is the maximum value in the original heat pump unit operation data, It is the normalized operating data of the heat pump unit.

3. The electronic expansion valve control method based on deep dynamic fuzzy neural network according to claim 1 is characterized in that: The step of dividing the normalized heat pump unit operation data into a training set and a test set further includes: The normalized heat pump unit operation data is divided into a training set and a test set according to a ratio of 7:

3.

4. The electronic expansion valve control method based on deep dynamic fuzzy neural network according to claim 1 is characterized in that: The deep dynamic fuzzy neural network model includes an input layer, a membership function layer, a T-norm layer, a normalization layer and an output layer; wherein, The input layer is the input port of data; The formula of the membership function layer is: in, is the normalized heat pump unit operation data, i = {1, 2, ..., r} is the amount of input data, j = {1, 2, ..., u} represents the number of rules, u ij and c ij are the membership functions and their centers, σ ij is the width of the Gaussian function at different layers; The T-norm layer, its j layer The formula is: The normalization layer, its j layer The formula is: The output layer has an output function y(x) as follows: Among them, w k is the weight of the kth layer, is the output function of the kth layer.

5. The electronic expansion valve control method based on deep dynamic fuzzy neural network according to claim 1 is characterized in that: The process of the genetic algorithm includes: S1.1: Encoding of membership function, using triangular membership function, with Z-type and S-type functions used at the beginning and end respectively, and using the intermediate parameters to be optimized as the triangle vertices to avoid gaps; S1.2: Initialize the population, set the number of intersections, population size and number of iterations; S1.3: Evaluate the fitness of individuals in the population. The fitness formula is: Among them, f I is the fitness function of the genetic algorithm, m i and l i is the value of the x-coordinate and the y-coordinate; the distance calculation formula between two adjacent nodes and three adjacent nodes is: Where: d 1i is the distance between any two adjacent nodes in the path, d 2i is the distance between any three adjacent nodes in the path, and n is the total number of nodes in the total path; S1.4: Based on the population fitness calculated from the initial population, the optimal individual set and the fitness set of each generation are obtained; S1.5: Select a two-point crossover operator to perform calculations to obtain a new generation of populations, and then mutate the new generation of populations to obtain a new generation of populations; S1.6: To avoid the loss of the best individual, the calculation results are stored, including each initial population, each iteration population information, and each iteration population individual information, so as to obtain the optimal parameters.

6. The electronic expansion valve control method based on deep dynamic fuzzy neural network according to claim 1 is characterized in that: The step of obtaining the heat pump unit operation data further includes: The operation data of the heat pump unit is collected through the wire controller and various sensors.

Images