Self-heat recovery method and system of compression heat pump based on neural network model

Abstract

The invention relates to the field of internet control, in particular to a self-heat recovery method and system for a compression heat pump based on a neural network model, and the method comprises the steps: inputting a flow path of a heat transfer medium; based on the flowing path, historical data of the heat transfer medium in a preset time period are collected; inputting the historical data into a preset neural network model, and training the neural network model; collecting real-time data of the heat transfer medium; on the basis of the trained neural network model, real-time data are analyzed, and predicted quantities are output and comprise predicted heat absorption capacity and predicted heat release capacity; based on different conditions, different heat recovery modes are triggered; when the real-time data is within the preset normal range, the heat recovery mode is stopped; and sending the feedback report to the user side. The self-heating recovery device has the effect that the self-heating recovery efficiency of the compression heat pump is higher.

Description

Technical Field

[0001] The present application relates to the field of Internet control, and specifically to a self-heat recovery method and system for a compression heat pump based on a neural network model. Background Art

[0002] With the continuous increase in energy consumption and the growing prominence of environmental problems, heat pump technology, as an efficient energy utilization method, has been widely used in the fields of refrigeration and heating. Compression heat pumps are favored due to their high thermal energy conversion efficiency. However, in practical applications, compression heat pumps have certain limitations in self-heat recovery, resulting in the energy utilization rate not reaching the optimal level. Existing heat pump systems usually lack an effective self-heat recovery mechanism and fail to fully utilize the waste heat during system operation.

[0003] Therefore, improvement is needed. Summary of the Invention

[0004] In view of the above problems, the present application proposes a self-heat recovery method and system for a compression heat pump based on a neural network model.

[0005] The first invention object of the present application is achieved through the following technical solutions.

[0006] A self-heat recovery method for a compression heat pump based on a neural network model includes the steps of:

[0007] Input the flow path of the heat transfer medium, where the flow path includes key components, and the key components include an endothermic heat exchanger, a compressor, an exothermic heat exchanger, and an expansion valve;

[0008] Based on the flow path, collect historical data of the heat transfer medium within a preset time period, where the historical data includes historical pressure data and historical temperature data;

[0009] Input the historical data into a preset neural network model and train the neural network model;

[0010] Collect real-time data of the heat transfer medium, where the real-time data includes real-time pressure data and real-time temperature data;

[0011] Based on the trained neural network model, analyze the real-time data and output a prediction quantity, where the prediction quantity includes predicted heat absorption and predicted heat release;

[0012] Trigger different heat recovery modes based on different situations;

[0013] When the real-time data is within a preset normal range, stop the heat recovery mode;

[0014] Send a feedback report to the user terminal.

[0015] In a preferred embodiment, the steps of inputting historical data into a preset neural network model and training the neural network model include the steps:

[0016] Preprocess the historical data, where the preprocessing includes data cleaning and normalization;

[0017] For the preprocessed historical data, calculate and extract key features, where the key features include average pressure, average temperature, pressure change rate, temperature change rate, and the correlation between pressure and temperature.

[0018] In a preferred embodiment, the steps of inputting historical data into a preset neural network model and training the neural network model further include the steps:

[0019] Determine the number of input layer nodes as the number of key features and the number of output layer nodes as 1;

[0020] Input the key features into the preset neural network model, and based on the gradient descent algorithm, update the network weights through multiple iterations, including the formula:

[0021]

[0022] w is the network weight, b is the bias, X is the input key feature, y is the true label, α is the learning rate, is the gradient of the error function with respect to the network weight.

[0023] In a preferred embodiment, the steps of analyzing real-time data based on the trained neural network model and outputting a prediction quantity, where the prediction quantity includes predicted heat absorption and predicted heat release, include the steps:

[0024] Input the real-time data into the trained neural network model;

[0025] Based on the preset formula σ is the activation function, and perform forward propagation;

[0026] Output the predicted heat absorption f1 is the first activation function;

[0027] Output the predicted heat release f2 is the second activation function.

[0028] In a preferred embodiment, the steps of triggering different heat recovery modes based on different situations include the steps:

[0029] When in the heat absorption process, When δ1 is the first preset threshold, trigger the efficient heat absorption recovery mode;

[0030] Based on the preset control strategy, increase the flow rate of the heat transfer medium;

[0031] Adjust the opening degree φ of the expansion valve, the rotational speed n of the compressor, and the suction volume V;

[0032] Based on a preset expansion valve opening adjustment amount Δφ1, reduce the opening degree φ of the expansion valve:

[0033] φ new = φ current -Δφ1;

[0034] Based on a preset compressor rotational speed adjustment amount Δn1, reduce the rotational speed n of the compressor: n new = n current -Δn1;

[0035] Based on a preset suction volume adjustment amount ΔV1, reduce the suction volume V:

[0036] V new = V current -ΔV1.

[0037] In a preferred embodiment, the step of triggering different heat recovery modes based on different situations further includes the steps:

[0038] When in the heat release process, δ2 is a second preset threshold, trigger the high-efficiency heat release recovery mode;

[0039] Based on a preset control strategy, reduce the flow rate of the heat transfer medium;

[0040] Adjust the opening degree φ of the expansion valve, the rotational speed n of the compressor, and the suction volume V;

[0041] Based on a preset expansion valve opening adjustment amount Δφ2, increase the opening degree φ of the expansion valve:

[0042] φ new = φ current +Δφ2;

[0043] Based on a preset compressor rotational speed adjustment amount Δn2, increase the rotational speed n of the compressor: n new = n current +Δn2;

[0044] Based on a preset suction volume adjustment amount ΔV2, increase the suction volume V:

[0045] V new = V current +ΔV2.

[0046] In a preferred embodiment, the step of stopping the heat recovery mode when the real-time data is within a preset normal range includes the steps:

[0047] Preset P real as the real-time pressure, Treal is the real-time temperature, P normal is the pressure in the normal working state, T real is the temperature range in the normal working state, δ P is the allowable pressure deviation range, δ T is the allowable temperature deviation range;

[0048] When P normal -δ P <P real <P normal +δ P , T normal -δ T <T real <T normal +δ T stop the heat recovery mode.

[0049] The second invention object of this application is achieved by the following technical solutions:

[0050] A compression heat pump self-heat recovery system based on a neural network model, comprising:

[0051] Key component module: input the flow path of the heat transfer medium, the flow path includes key components, and the key components include an endothermic heat exchanger, a compressor, an exothermic heat exchanger, and an expansion valve;

[0052] Collection module: based on the flow path, collect historical data of the heat transfer medium within a preset time period, and the historical data includes historical pressure data and historical temperature data;

[0053] Training module: input the historical data into a preset neural network model and train the neural network model;

[0054] Acquisition module: acquire real-time data of the heat transfer medium, and the real-time data includes real-time pressure data and real-time temperature data;

[0055] Output module: based on the trained neural network model, analyze the real-time data and output predictions, and the predictions include predicted heat absorption and predicted heat release;

[0056] Trigger module: trigger different heat recovery modes based on different situations;

[0057] Stop module: stop the heat recovery mode when the real-time data is within the preset normal range;

[0058] Sending module: send a feedback report to the user side.

[0059] The third object of this application is achieved by the following technical solutions:

[0060] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned compression heat pump self-heat recovery method based on a neural network model are implemented.

[0061] The above-mentioned fourth object of the present application is achieved by the following technical solutions:

[0062] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned compression heat pump self-heat recovery method based on a neural network model are implemented.

[0063] In summary, the present application includes at least one of the following beneficial technical effects:

[0064] 1. By inputting the flow path of the heat transfer medium and key component information, collecting and analyzing historical and real-time data, training a neural network model to predict the heat absorption and heat release, and thus triggering different heat recovery modes according to real-time data, stopping heat recovery when the heat pump operates within the normal range to optimize energy utilization efficiency, and enabling users to monitor and adjust the system by sending feedback reports, clearly and logically solving the problem of insufficient waste heat utilization in the prior art.

[0065] 2. By preprocessing historical data, including data cleaning and normalization, to ensure data consistency and accuracy; then calculating and extracting key features from the preprocessed data, such as average pressure, average temperature, etc., these features help to reveal the heat energy conversion law during the operation of the heat pump, thus providing a basis for subsequent optimization of the self-heat recovery mechanism, specifically and clearly solving the problem that the existing heat pump system fails to fully utilize waste heat.

[0066] 3. By setting the number of nodes in the input layer of the neural network model to the number of key features and the number of nodes in the output layer to 1, so that the model can effectively learn based on these key features; then inputting the key features into the neural network model and using the gradient descent algorithm to update the network weights through multiple iterations, where the formula describes the process of weight update, including network weights, biases, input features, true labels, learning rates, and error gradients. This logical process is specifically and clearly aimed at training a model that can accurately predict the heat recovery efficiency, thus optimizing the self-heat recovery mechanism of the heat pump system and improving energy utilization efficiency.

[0067] 4. First, input the real-time data into the already trained neural network model; then perform forward propagation calculation using the preset activation function, which is the core process of the neural network for processing data; next, output the predicted heat absorption and heat release amounts respectively, and different activation functions are used here to obtain these two key heat recovery parameters. This series of logical steps specifically and clearly solves the problem that the existing heat pump system fails to make full use of waste heat, optimizes the self-heat recovery efficiency through the prediction ability of the neural network model, and thus improves the overall energy utilization rate. Description of the Drawings

[0068] Figure 1 is a flowchart of an implementation of an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0069] Figure 2 is a flowchart of an implementation of step S30 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0070] Figure 3 is another flowchart of an implementation of step S30 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0071] Figure 4 is a flowchart of an implementation of step S50 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0072] Figure 5 is a flowchart of an implementation of step S60 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0073] Figure 6 is another flowchart of an implementation of step S60 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0074] Figure 7 is a flowchart of an implementation of step S70 in an embodiment of a compression heat pump self-heat recovery method based on a neural network model in the present application;

[0075] Figure 8 is a principle block diagram of a computer device in the present application. Detailed Embodiments

[0076] The following is a further detailed description of the present application in conjunction with the attached Figure 1-8 drawings.

[0077] In one embodiment, as Figure 1 shown, the present application discloses a compression heat pump self-heat recovery method based on a neural network model, which specifically includes the following steps:

[0078] S10: Input the flow path of the heat transfer medium, where the flow path includes key components, and the key components include an endothermic heat exchanger, a compressor, an exothermic heat exchanger, and an expansion valve;

[0079] S20: Based on the flow path, collect the historical data of the heat transfer medium within a preset time period, where the historical data includes historical pressure data and historical temperature data;

[0080] S30: Input the historical data into a preset neural network model and train the neural network model;

[0081] S40: Collect the real-time data of the heat transfer medium, where the real-time data includes real-time pressure data and real-time temperature data;

[0082] S50: Based on the trained neural network model, analyze the real-time data and output a prediction quantity, where the prediction quantity includes predicted heat absorption and predicted heat release;

[0083] S60: Trigger different heat recovery modes based on different situations;

[0084] S70: When the real-time data is within a preset normal range, stop the heat recovery mode;

[0085] S80: Send a feedback report to the user terminal.

[0086] In this embodiment, by inputting the flow path of the heat transfer medium and key component information, collecting and analyzing historical and real-time data, training a neural network model to predict heat absorption and heat release, thereby triggering different heat recovery modes according to real-time data, stopping heat recovery when the heat pump operates within a normal range to optimize energy utilization rate, and enabling users to monitor and adjust the system by sending a feedback report, the problem of insufficient waste heat utilization in the prior art is clearly solved logically.

[0087] Figure 2 , Step S30 includes the steps of:

[0088] S301: Preprocess the historical data, where the preprocessing includes data cleaning and normalization;

[0089] S302: For the preprocessed historical data, calculate and extract key features, where the key features include average pressure, average temperature, pressure change rate, temperature change rate, and correlation between pressure and temperature.

[0090] In this embodiment, the historical data is preprocessed in step S301, including data cleaning and normalization, to ensure data consistency and accuracy; then in step S302, key features such as average pressure and average temperature are calculated and extracted from the preprocessed data. These features help to reveal the heat energy conversion law during the operation of the heat pump, thereby providing a basis for subsequent optimization of the self-heat recovery mechanism, specifically and clearly solving the problem that the existing heat pump system fails to fully utilize waste heat.

[0091] Figure 3 , step S30 also includes the steps:

[0092] S303: Determine the number of input layer nodes as the number of key features, and the number of output layer nodes as 1;

[0093] S304: Input the key features into a preset neural network model, and based on the gradient descent algorithm, update the network weights through multiple iterations, including the formula:

[0094] w new = w old -α·▽ w Loss(w,b,X,y);

[0095] S305: w is the network weight, b is the bias, X is the input key feature, y is the true label, α is the learning rate, and ▽ w Loss is the gradient of the error function with respect to the network weight.

[0096] In this embodiment, in step S303, the number of input layer nodes of the neural network model is set as the number of key features, and the number of output layer nodes is set as 1, so that the model can effectively learn based on these key features; then in step S304, the key features are input into the neural network model, and the gradient descent algorithm is used to update the network weights through multiple iterations. The formula describes the process of weight update, including network weight, bias, input feature, true label, learning rate, and error gradient. This logical process is specifically and clearly aimed at training a model that can accurately predict the heat recovery efficiency, thereby optimizing the self-heat recovery mechanism of the heat pump system and improving the energy utilization rate.

[0097] Figure 4 , step S50 includes the steps:

[0098] S501: Input the real-time data into the trained neural network model;

[0099] S502: Based on the preset formula σ is the activation function, and perform forward propagation;

[0100] S503: Output the predicted heat absorption f1 is the first activation function;

[0101] S504: Output the predicted heat release f2 is the second activation function.

[0102] In this embodiment, first in step S501, the real-time data is input into the already trained neural network model; then in step S502, forward propagation calculation is performed using a preset activation function, which is the core of the neural network's data processing; then in steps S503 and S504, the predicted heat absorption and heat release are output respectively, and different activation functions are used here to obtain these two key heat recovery parameters. This series of logical steps specifically and clearly solves the problem that the existing heat pump system fails to fully utilize waste heat, optimizes the self-heat recovery efficiency through the prediction ability of the neural network model, and thus improves the overall energy utilization rate.

[0103] Figure 5 , step S60 includes the steps:

[0104] S601: When in the heat absorption process, when δ1 is the first preset threshold, trigger the high-efficiency heat absorption recovery mode;

[0105] S602: Based on the preset control strategy, increase the flow rate of the heat transfer medium;

[0106] S603: Adjust the expansion valve opening φ, the compressor speed n, and the suction volume V;

[0107] S604: Based on the preset expansion valve opening adjustment amount Δφ1, reduce the expansion valve opening φ:

[0108] φ new = φ current -Δφ1;

[0109] S605: Based on the preset compressor speed adjustment amount Δn1, reduce the compressor speed n:

[0110] n new = n current -Δn1;

[0111] S606: Based on the preset suction volume adjustment amount ΔV1, reduce the suction volume V:

[0112] V new = V current -ΔV1.

[0113] In this embodiment, in step S601, when it is detected that the heat absorption amount exceeds the first preset threshold, the high-efficiency heat absorption recovery mode is triggered; then in step S602, the flow rate of the heat transfer medium is increased according to the preset control strategy to improve the heat recovery efficiency; then in step S603, the heat recovery process is further optimized by adjusting the expansion valve opening, the compressor speed, and the suction volume; specifically in steps S604, S605, and S606, the expansion valve opening, the compressor speed, and the suction volume are reduced respectively according to the preset adjustment amounts, and these adjustments are aimed at precisely controlling the operation of the heat pump to improve the self-heat recovery efficiency. This series of logical steps specifically and clearly solves the problem that the existing heat pump system fails to make full use of waste heat, thereby optimizing the energy utilization rate.

[0114] Figure 6 , step S60 also includes the steps:

[0115] SA1: When during the heat release process, δ2 is the second preset threshold, the high-efficiency heat release recovery mode is triggered;

[0116] SA2: Based on the preset control strategy, reduce the flow rate of the heat transfer medium;

[0117] SA3: Adjust the expansion valve opening φ, the compressor speed n, and the suction volume V;

[0118] SA4: Based on the preset expansion valve opening adjustment amount Δφ2, increase the expansion valve opening φ:

[0119] φ new = φ current +Δφ2;

[0120] SA5: Based on the preset compressor speed adjustment amount Δn2, increase the compressor speed n:

[0121] n new = n current +Δn2;

[0122] SA6: Based on the preset suction volume adjustment amount ΔV2, increase the suction volume V:

[0123] V new = V current +ΔV2.

[0124] In this embodiment, in step SA1, when a certain parameter during the heat release process reaches the second preset threshold, the system will trigger the high-efficiency heat release recovery mode; in order to recover heat more effectively, in step SA2, the flow rate of the heat transfer medium is reduced according to a preset control strategy; then in step SA3, the heat release efficiency of the heat pump is optimized by adjusting the expansion valve opening, compressor speed, and suction volume; specifically in steps SA4, SA5, and SA6, the expansion valve opening, compressor speed, and suction volume are increased respectively according to the preset adjustment amounts. These adjustments help to improve the performance of the heat pump in the heat release mode, thereby making more full use of waste heat, clearly solving the problem that the existing heat pump system fails to make full use of waste heat, and improving the overall energy utilization rate.

[0125] Figure 7 , step S70 includes the steps:

[0126] S701: Preset P real as the real-time pressure, T real as the real-time temperature, P normal as the pressure in the normal working state, T real as the temperature range in the normal working state, δ P as the allowable pressure deviation range, δ T as the allowable temperature deviation range;

[0127] S702: When P normal -δ P <P real <P normal +δ P , T normal -δ T <T real <T normal +δ T , stop the heat recovery mode.

[0128] In this embodiment, in step S701, parameters of the real-time pressure and temperature, as well as the pressure and temperature ranges in the normal working state are preset, and at the same time, the allowable pressure and temperature deviation ranges are set; in step S702, when the real-time monitored pressure exceeds the pressure range in the normal working state plus the allowable pressure deviation range, or the real-time monitored temperature exceeds the temperature range in the normal working state plus the allowable temperature deviation range, the system will stop the heat recovery mode. This logical process specifically and clearly ensures that the heat recovery can be stopped in time when the heat pump operates abnormally, preventing further energy consumption losses, thereby protecting the heat pump system and optimizing its energy utilization rate.

[0129] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0130] In one embodiment, a compression heat pump self-heat recovery system based on a neural network model is provided. This compression heat pump self-heat recovery system based on a neural network model corresponds to a compression heat pump self-heat recovery method based on a neural network model in the above embodiment. This compression heat pump self-heat recovery system based on a neural network model includes:

[0131] Key component module: input the flow path of the heat transfer medium, the flow path includes key components, and the key components include an endothermic heat exchanger, a compressor, an exothermic heat exchanger, and an expansion valve;

[0132] Collection module: based on the flow path, collect historical data of the heat transfer medium within a preset time period, and the historical data includes historical pressure data and historical temperature data;

[0133] Training module: input the historical data into a preset neural network model and train the neural network model;

[0134] Acquisition module: collect real-time data of the heat transfer medium, and the real-time data includes real-time pressure data and real-time temperature data;

[0135] Output module: based on the trained neural network model, analyze the real-time data and output a prediction quantity, and the prediction quantity includes predicted heat absorption and predicted heat release;

[0136] Trigger module: trigger different heat recovery modes based on different situations

[0137] Stop module: when the real-time data is within a preset normal range, stop the heat recovery mode;

[0138] Sending module: send a feedback report to the user side.

[0139] Optionally, it further includes:

[0140] Preprocessing module: preprocess the historical data, and the preprocessing includes data cleaning and normalization;

[0141] Calculation and extraction module: calculate and extract key features from the preprocessed historical data, and the key features include average pressure, average temperature, pressure change rate, temperature change rate, and correlation between pressure and temperature.

[0142] Optionally, it further includes:

[0143] Determination module: determine the number of input layer nodes as the number of key features and the number of output layer nodes as 1;

[0144] Update module: input the key features into a preset neural network model, and based on the gradient descent algorithm, update the network weights through multiple iterations, including the formula:

[0145]

[0146] The first definition module: w is the network weight, b is the bias, X is the input key feature, y is the true label, α is the learning rate, is the gradient of the error function with respect to the network weight.

[0147] Optionally, it further includes:

[0148] The input module: Inputs real-time data into the trained neural network model;

[0149] The forward propagation module: Based on the preset formula where σ is the activation function, performs forward propagation;

[0150] The first output module: Outputs the predicted heat absorption f1 is the first activation function;

[0151] The second output module: Outputs the predicted heat release f2 is the second activation function.

[0152] Optionally, it further includes:

[0153] The first trigger module: When during the heat absorption process, where δ1 is the first preset threshold, triggers the efficient heat absorption recovery mode;

[0154] The flow rate increasing module: Increases the flow rate of the heat transfer medium based on the preset control strategy;

[0155] The first adjustment module: Adjusts the expansion valve opening φ, the compressor speed n, and the suction volume V;

[0156] The first reduction module: Reduces the expansion valve opening φ based on the preset expansion valve opening adjustment amount Δφ1:

[0157] φ new = φ current - Δφ1;

[0158] The second reduction module: Reduces the compressor speed n based on the preset compressor speed adjustment amount Δn1:

[0159] n new = n current - Δn1;

[0160] The third reduction module: Reduces the suction volume V based on the preset suction volume adjustment amount ΔV1:

[0161] V new = V current - ΔV1.

[0162] Optionally, it further includes:

[0163] The second trigger module: When during the heat release process, δ2 is the second preset threshold, trigger the high-efficiency heat release recovery mode;

[0164] The flow reduction module: Based on a preset control strategy, reduce the flow rate of the heat transfer medium;

[0165] The second adjustment module: Adjust the expansion valve opening φ, the compressor speed n, and the suction volume V;

[0166] The first increase module: Based on a preset expansion valve opening adjustment amount Δφ2, increase the expansion valve opening φ:

[0167] φ new = φ current + Δφ2;

[0168] The second increase module: Based on a preset compressor speed adjustment amount Δn2, increase the compressor speed n:

[0169] n new = n current + Δn2;

[0170] The third increase module: Based on a preset suction volume adjustment amount ΔV2, increase the suction volume V:

[0171] V new = V current + ΔV2.

[0172] Optionally, it further includes:

[0173] The second definition module: Preset P real as the real-time pressure, T real as the real-time temperature, P normal as the pressure in the normal working state, T real as the temperature range in the normal working state, δ P as the allowable pressure deviation range, δ T as the allowable temperature deviation range;

[0174] The heat recovery stop module: When P normal - δ P < P real < P normal + δ P , T normal - δ T < T real < T normal + δ T , stop the heat recovery mode.

[0175] For the specific limitations of a compression heat pump self-heat recovery system based on a neural network model, reference can be made to the limitations of a compression heat pump self-heat recovery method based on a neural network model in the above text, which will not be elaborated here. Each module in the above compression heat pump self-heat recovery system based on a neural network model can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0176] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 8 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store deviations. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a compression heat pump self-heat recovery method based on a neural network model.

[0177] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it is a compression heat pump self-heat recovery method based on a neural network model.

[0178] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, it is a compression heat pump self-heat recovery method based on a neural network model.

[0179] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0180] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

Claims

1. A compression heat pump self-heat recovery method based on a neural network model, characterized in that: Includes steps: A flow path for inputting a heat transfer medium, wherein the flow path includes key components, including a heat absorbing heat exchanger, a compressor, a heat releasing heat exchanger, and an expansion valve; Based on the flow path, collecting historical data of the heat transfer medium within a preset time period, the historical data including historical pressure data and historical temperature data; Input historical data into a preset neural network model and train the neural network model; Collecting real-time data of the heat transfer medium, wherein the real-time data includes real-time pressure data and real-time temperature data; Based on the trained neural network model, real-time data is analyzed and predicted quantities are output, wherein the predicted quantities include predicted heat absorption and predicted heat release; Trigger different heat recovery modes based on different situations; When the real-time data is within the preset normal range, the heat recovery mode is stopped; Send feedback report to the user end.

2. A compression heat pump self-heat recovery method based on a neural network model according to claim 1, characterized in that: The step of inputting historical data into a preset neural network model and training the neural network model comprises the steps of: Preprocessing historical data, including data cleaning and normalization; The key features are calculated and extracted from the preprocessed historical data, and the key features include average pressure, average temperature, pressure change rate, temperature change rate, and correlation between pressure and temperature.

3. The method for self-heat recovery of a compression heat pump based on a neural network model according to claim 1, characterized in that: The step of inputting historical data into a preset neural network model and training the neural network model further includes the steps of: The number of input layer nodes is determined as the number of key features, and the number of output layer nodes is determined as 1; Input key features into the preset neural network model, and update the network weights through multiple iterations based on the gradient descent algorithm, including the formula: w new =w old -α·▽ w Loss(w,b,X,y); w is the network weight, b is the bias, X is the key feature of the input, y is the true label, α is the learning rate, ▽ w Loss is the gradient of the error function with respect to the network weights.

4. The method for self-heat recovery of a compression heat pump based on a neural network model according to claim 1, characterized in that: The method is based on the trained neural network model, analyzing real-time data and outputting a predicted amount, wherein the predicted amount includes the steps of predicting heat absorption and heat release, including the steps of: Input real-time data into the trained neural network model; Based on preset formula σ is the activation function for forward propagation; Output predicted heat absorption f1 is the first activation function; Output predicted heat release f2 is the second activation function.

5. The method for self-heat recovery of a compression heat pump based on a neural network model according to claim 1, characterized in that: The steps of triggering different heat recovery modes based on different situations include the following steps: During the endothermic process, When δ1 is the first preset threshold, the high-efficiency heat absorption recovery mode is triggered; Based on the preset control strategy, increase the flow rate of the heat transfer medium; Adjust the expansion valve opening φ, compressor speed n, and suction volume V; Based on the preset expansion valve opening adjustment amount Δφ1, reduce the expansion valve opening φ: f new =φ current -Df1; Based on the preset compressor speed adjustment amount Δn1, reduce the compressor speed n: n new =n current -Δn1; Based on the preset air intake adjustment amount ΔV1, reduce the air intake volume V: V new =V current -ΔV1。 6. The method for self-heat recovery of a compression heat pump based on a neural network model according to claim 1, characterized in that: The step of triggering different heat recovery modes based on different situations further includes the steps of: During the heat release process, When δ2 is the second preset threshold, the high-efficiency heat release recovery mode is triggered; Based on the preset control strategy, reduce the flow rate of heat transfer medium; Adjust the expansion valve opening φ, compressor speed n, and suction volume V; Based on the preset expansion valve opening adjustment amount Δφ2, increase the expansion valve opening φ: f new =φ current +Δφ2; Based on the preset compressor speed adjustment amount Δn2, the compressor speed n is increased: n new =n current +Δn2; Based on the preset air intake adjustment amount ΔV2, increase the air intake volume V: V new =V current +ΔV2。 7. The method for self-heat recovery of a compression heat pump based on a neural network model according to claim 1, characterized in that: The step of stopping the heat recovery mode when the real-time data is within a preset normal range comprises the following steps: Preset P real is the real-time pressure, T real is the real-time temperature, P normal is the pressure in normal working state, T real is the temperature range of normal working condition, δ P is the allowable pressure deviation range, δ T is the allowable temperature deviation range; When P normal -δ P <P real <P normal +δ P , T normal -δ T <T real <T normal +δ T When the heat recovery mode is stopped.

8. A compression heat pump self-heat recovery system based on a neural network model, comprising: Key component module: a flow path for inputting heat transfer medium, the flow path includes key components, the key components include a heat absorbing heat exchanger, a compressor, a heat releasing heat exchanger, and an expansion valve; Collection module: based on the flow path, collects historical data of the heat transfer medium within a preset time period, the historical data including historical pressure data and historical temperature data; Training module: input historical data into the preset neural network model and train the neural network model; Acquisition module: collects real-time data of heat transfer medium, including real-time pressure data and real-time temperature data; Output module: Based on the trained neural network model, analyze the real-time data and output the predicted quantity, which includes the predicted heat absorption and the predicted heat release; Trigger module: trigger different heat recovery modes based on different situations Stop module: When the real-time data is within the preset normal range, the heat recovery mode is stopped; Sending module: Send feedback report to the user end.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the compression heat pump self-heat recovery method based on the neural network model as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of a compression heat pump self-heat recovery method based on a neural network model as described in any one of claims 1 to 7 are implemented.

Images