Control method and device of air source heat pump unit, storage medium and server

Abstract

The application provides a control method and device of an air source heat pump unit, a storage medium and a server. The method comprises the following steps: obtaining an environmental parameter of a target region at a next time; inputting the obtained environmental parameter into a pre-trained n energy consumption optimization target model to respectively output n sets of control parameters; respectively distributing the n sets of control parameters to the n sets of air source heat pump units to respectively control the operation of the n sets of air source heat pump units; wherein the n sets of control parameters are respectively used for controlling the n sets of air source heat pump units at the next time; each set of air source heat pump units in the n sets of air source heat pump units comprises two or more air source heat pump units with similar conditions; and the n energy consumption optimization target models are trained for the n sets of air source heat pump units respectively. The scheme provided by the application can effectively solve the problems of data islanding and repeated training of a model.

Description

Technical Field

[0001] This invention relates to the field of control, and more particularly to a control method, apparatus, storage medium, and server for an air source heat pump unit. Background Technology

[0002] An air source heat pump heating system is an air conditioning and heating system that uses outdoor air as a heat source to produce hot air or hot water. It utilizes the ubiquitous ambient air as a low-grade energy source, using a small amount of electricity to drive a compressor and transfer heat. It eliminates the need for complex and expensive water intake, reinjection, or soil heat exchange systems and dedicated machine rooms, thus reducing the large amounts of pollutant emissions to the atmosphere caused by traditional heating methods. It ensures heating efficiency while being energy-saving and environmentally friendly. Compared to the three traditional methods of hot water production—electric heating, gas heating, and coal-fired water heating—air source heat pumps have significant advantages in both economy and environmental protection.

[0003] In related technologies, the energy-saving optimization direction for all air source heat pump units in a certain region is simultaneously distributed. However, because the data of different manufacturers is stored on their own servers and not shared due to privacy and other reasons, the data silos of air source heat pump operation from different manufacturers or the same manufacturer in different regions cannot be effectively utilized to address the problem of data silos. New technologies such as big data and artificial intelligence cannot be effectively utilized. Summary of the Invention

[0004] The main objective of this invention is to overcome the deficiencies of the aforementioned related technologies and provide a control method, device, storage medium, and server for air source heat pump units, so as to solve the problem of data silos in the operation of air source heat pumps from different manufacturers or from the same manufacturer in different regions.

[0005] This invention provides a control method for an air source heat pump unit, comprising: acquiring environmental parameters of a target area at the next moment; inputting the acquired environmental parameters into n pre-trained energy consumption optimization target models, and outputting n sets of control parameters respectively; distributing the n sets of control parameters to the n air source heat pump units respectively to control the operation of the n air source heat pump units respectively; wherein, the n sets of control parameters are used to control the n air source heat pump units at the next moment; each of the n air source heat pump units has similar conditions; and the n energy consumption optimization target models are trained for the n air source heat pump units respectively.

[0006] Optionally, the n energy consumption optimization target models are trained in the following manner: All air source heat pump units in the target area are divided into n groups of air source heat pump units with different similarity conditions; a shared model is trained for each of the n groups of air source heat pump units to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units; the n shared energy consumption models are used as the objective functions of the energy consumption optimization target models of the n groups of air source heat pump units, and energy consumption optimization target models corresponding to each of the n groups of air source heat pump units are built accordingly.

[0007] Optionally, shared model training is performed on each of the n groups of air source heat pump units to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units. This includes: the operating data for training the shared energy consumption model for each of the n groups of air source heat pump units is provided by k data providers; the k data providers encrypt the operating data for training the shared energy consumption model using a preset encryption tool and then send it to a trusted execution environment created by a preset computing platform for storage; the preset computing platform decrypts the encrypted operating data provided by the k data providers and then performs shared model training to obtain the corresponding shared energy consumption model.

[0008] Optionally, the acquired environmental parameters are input into n pre-trained energy consumption optimization target models, and n sets of control parameters are output respectively. This includes: after inputting the acquired environmental parameters into the n energy consumption optimization target models, using the preset control parameter range as a constraint, and solving them respectively using a heuristic algorithm to obtain the optimal solution of the n sets of control parameters.

[0009] Optionally, the environmental parameters include: outdoor temperature and / or solar radiation; and / or, the control parameters include: compressor frequency, outlet water temperature and / or indoor temperature.

[0010] Another aspect of the present invention provides a control device for an air source heat pump unit, comprising: an acquisition unit for acquiring environmental parameters of a target area at the next moment; an output unit for inputting the environmental parameters acquired by the acquisition unit into n pre-trained energy consumption optimization target models, and outputting n sets of control parameters respectively; and a distribution unit for distributing the n sets of control parameters to the n air source heat pump units respectively, so as to control the operation of the n air source heat pump units respectively; wherein, the n sets of control parameters are used to control the n air source heat pump units at the next moment; each of the n air source heat pump units has similar conditions; and the n energy consumption optimization target models are trained for the n air source heat pump units respectively.

[0011] Optionally, it further includes: a training unit, used to train the n energy consumption optimization target models in the following manner: dividing all air source heat pump units in the target area into n groups of air source heat pump units with different similarity conditions; training a shared model for each of the n groups of air source heat pump units to obtain n shared energy consumption models corresponding to the n groups of air source heat pump units; using the n shared energy consumption models as the objective functions of the energy consumption optimization target models of the n groups of air source heat pump units, and building the energy consumption optimization target models corresponding to each of the n groups of air source heat pump units.

[0012] Optionally, the training unit performs shared model training for each of the n groups of air source heat pump units to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units. This includes: the operating data for training the shared energy consumption model for each of the n groups of air source heat pump units is provided by k data providers; the k data providers encrypt the provided operating data for training the shared energy consumption model using a preset encryption tool and then send it to a trusted execution environment created by a preset computing platform for storage; the preset computing platform decrypts the encrypted operating data provided by the k data providers for training the shared energy consumption model and then performs shared model training to obtain the corresponding shared energy consumption model.

[0013] Optionally, the output unit inputs the acquired environmental parameters into n pre-trained energy consumption optimization target models and outputs n sets of control parameters respectively, including: after inputting the acquired environmental parameters into the n energy consumption optimization target models, using the preset control parameter range as a constraint condition, and solving them respectively using a heuristic algorithm to obtain the optimal solution of the n sets of control parameters.

[0014] Optionally, the environmental parameters include: outdoor temperature and / or solar radiation; and / or, the control parameters include: compressor frequency, outlet water temperature and / or indoor temperature.

[0015] In another aspect, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0016] In another aspect, the present invention provides a server including a processor, a memory, and a computer program stored in the memory that can run on the processor, wherein the processor executes the program to implement the steps of any of the methods described above.

[0017] In another aspect, the present invention provides a server including any of the control devices described above.

[0018] According to the technical solution of the present invention, energy-saving optimization control parameters are obtained by using a common energy consumption optimization model for air source heat pump units with similar conditions. All units with similar conditions are trained together to form a model, which can effectively cope with small sample scenarios where operating data is limited in some regions or manufacturers.

[0019] According to the technical solution of the present invention, the shared machine learning construction method based on TEE obtains the objective function of the unit energy consumption optimization target model, effectively solving the problems of data silos and repeated model training. The entire calculation process is safe and reliable. By establishing a shared energy consumption model of air source heat pump units under the same conditions, the optimal control variables are solved using heuristic algorithms, which efficiently guides the energy-saving direction of the units, helps to improve user experience, save energy consumption, and at the same time reduces the computing pressure on the server and the problem of insufficient training data when modeling a single unit.

[0020] According to the technical solution of the present invention, it can replace a single optimization strategy for a single unit, use a shared machine learning modeling method based on TEE to solve the server operating pressure problem, and efficiently issue energy-saving optimization instructions to a large number of units. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0022] Figure 1 This is a schematic diagram of an embodiment of the control method for an air source heat pump unit provided by the present invention;

[0023] Figure 2 A schematic diagram of a specific implementation method for training n groups of air source heat pump units to obtain n energy consumption optimization target models corresponding to each of the n groups of air source heat pump units is shown.

[0024] Figure 3 The training process of the shared model for each group of air source heat pump units with similar conditions is shown;

[0025] Figure 4 This is a schematic diagram of a specific embodiment of the energy-saving control method for air source heat pump units provided by the present invention;

[0026] Figure 5 An energy-saving optimization process for air-source heat pump units based on shared machine learning according to the present invention is shown.

[0027] Figure 6 This is a structural block diagram of an embodiment of the control device for an air source heat pump unit provided by the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] Operating data from air source heat pump units from different manufacturers or even the same manufacturer but in different regions may be stored on different servers, creating data silos due to security concerns; or the collected data may be small (i.e., insufficient data volume), making it impossible to establish an effective energy consumption model.

[0031] This invention provides a control method for an air source heat pump unit.

[0032] Figure 1 This is a schematic diagram of an embodiment of the control method for an air source heat pump unit provided by the present invention.

[0033] like Figure 1 As shown, according to an embodiment of the present invention, the control method of the air source heat pump unit includes at least steps S110, S120 and S130.

[0034] Step S110: Obtain the environmental parameters of the target area at the next moment.

[0035] The environmental parameters may specifically include outdoor temperature and / or solar radiation. The outdoor temperature and / or solar radiation can be obtained using web crawling technology, for example, by crawling weather forecast websites to obtain the outdoor temperature and / or solar radiation for the next moment.

[0036] Step S120: Input the acquired environmental parameters into n pre-trained energy consumption optimization target models, and output n sets of control parameters respectively.

[0037] The n sets of control parameters are used to control the n air source heat pump units at the next time step. Each of the n air source heat pump units includes two or more air source heat pump units with similar conditions. The n energy consumption optimization target models are trained on the n air source heat pump units respectively. That is, each group of air source heat pump units with similar conditions has a corresponding energy consumption optimization target model.

[0038] Figure 2 This diagram illustrates a specific implementation process for training n groups of air-source heat pump units to obtain n energy consumption optimization target models. (Example:) Figure 2 As shown, in one specific implementation, the n energy consumption optimization target models are trained in the following manner;

[0039] Step S121: Divide all air source heat pump units in the target area into n groups of air source heat pump units with different similar conditions.

[0040] In one specific implementation, all air source heat pump units in the target area are divided into n groups of air source heat pump units with different similar conditions based on the unit's performance data, installation data, and local climate data.

[0041] The installation data may specifically include: heat source installation location and / or terminal type. For example, the similarity condition classification of air source heat pump units needs to consider factors such as: unit performance, climate zone, heat source installation location, and terminal. The first factor is unit performance; units with different performance characteristics have significant differences in energy consumption models. The second factor is the climate zone where the unit is located; units in extremely cold climate zones have higher energy consumption. The third factor is the heat source installation location; different locations, such as balconies, rooftops, and basements, are related to airflow. The fourth factor is the terminal, such as underfloor heating, fan coil units, or radiators, which affects the unit's heating effect. For example, two units can be considered similar air source heat pump units only if they have the same performance, are located in the same climate zone, have comparable airflow at the heat source, and have the same terminal. Each of the n groups of air source heat pump units has similar conditions. Each of the n groups of air source heat pump units includes two or more air source heat pump units with similar conditions.

[0042] Step S122: For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to the n groups of air source heat pump units.

[0043] Specifically, all air source heat pump units are divided into n groups of similar-condition devices (i.e., n groups of air source heat pump units with similar conditions). Data for each group of similar-condition devices (i.e., each group of air source heat pump units) may be stored on k different servers, meaning that data for each group of similar-condition devices (including two or more devices) may have k data participants. Specifically, the operational data for training the shared energy consumption model for each of the n groups of air source heat pump units is provided by k data providers. Each of the k data providers encrypts its provided operational data using a preset encryption tool and sends it to a trusted execution environment created by a preset computing platform for storage. The preset computing platform decrypts the encrypted operational data provided by the k data providers and performs shared model training to obtain the corresponding shared energy consumption model. Shared model training is then performed for each of the n groups of air source heat pump units, resulting in n shared energy consumption models corresponding to each of the n groups of air source heat pump units.

[0044] To address the issue of data silos in the operation of air source heat pumps from different manufacturers or even the same manufacturer in different regions, this invention employs a TEE-based shared model to obtain the objective function of the optimization model (energy consumption optimization target model). The optimization model is used for energy-saving optimization of computer groups (i.e., solving the corresponding optimization model).

[0045] For details on training the shared model, please refer to [link / reference]. Figure 3The training process of the shared model for each group of air source heat pump units with similar conditions is shown.

[0046] like Figure 3 As shown, the training process of the shared model for each group of devices with similar conditions includes:

[0047] S31, k data providers download and deploy encryption tools from the computing platform.

[0048] Due to privacy concerns regarding data (operational data and installation data), data from a single similar device exists on k servers belonging to k data participants and cannot be transferred between them.

[0049] S32, k data providers encrypt the data using encryption tools;

[0050] S33, the computing platform creates a trusted execution environment (TEE);

[0051] S34, k data providers send encrypted data to a trusted execution environment storage;

[0052] S35, The task initiator initiates the computation task;

[0053] S36, decrypt k data items and verify the authorization process;

[0054] S37. The computing platform trains the model based on the decrypted data to obtain the shared energy consumption model of the similar device, that is, the objective function of the optimization model corresponding to the similar device.

[0055] S38, transmits the encrypted calculation result;

[0056] S39, Destroy the trusted execution environment. (The destruction step is performed after the optimization model is solved; otherwise, the shared model cannot participate in the optimization model solution calculation.)

[0057] For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units. This model is used to predict the energy consumption of air source heat pump units with similar conditions. The training data for the shared energy consumption models selects features related to the energy consumption of the air source heat pump units as model inputs. The models are trained based on energy consumption, and the training data may include, for example, at least one of indoor temperature, outdoor temperature, compressor frequency, outlet water temperature, solar radiation, and power consumption.

[0058] Data providers encrypt and send data to a trusted computing platform. The platform decrypts the data and trains a model (which can be a relevant AI regression algorithm). Once certain conditions are met (reaching the required number of iterations or the error is less than a threshold), model training is complete, resulting in a shared energy consumption model for similar devices. Note that due to encryption, the trained model is not readily available. Data owners or others authorizing access to the shared model can input information and obtain return values ​​using an Application Programming Interface (API).

[0059] Step S123: Use the n shared energy consumption models as the objective functions of the energy consumption optimization target models of the n groups of air source heat pump units, and build the energy consumption optimization target model corresponding to each group of air source heat pump units in the n groups of air source heat pump units.

[0060] Due to encryption, the trained shared energy consumption model cannot be obtained. When the data owner or others authorize access to the shared energy consumption model, they can input information and obtain return values ​​using the Application Programming Interface (API). Let the encrypted shared energy consumption model be denoted as E. Since n shared energy consumption models (i.e., n objective functions) have been trained for each of the n groups of air-source heat pump units, n optimization models (i.e., n energy consumption optimization objective models) can be built accordingly. The complete energy consumption optimization objective models are shown below:

[0061]

[0062] Where E represents the shared energy consumption model, T i Indicates indoor temperature, T o Indicates outdoor temperature, f com T represents the compressor frequency. out Indicates the outlet water temperature, r b This represents the amount of solar radiation. The energy consumption optimization objective model limits the outlet water temperature T. out Compressor frequency f com Indoor temperature T i The control parameter range. Outdoor temperature To and solar radiation rb can be collected via web crawler. and The outlet water temperature T represents the water temperature T. out The maximum and minimum values; and These represent the compressor frequency f. com The maximum and minimum values, and These represent the indoor temperature T. i The maximum and minimum values.

[0063] Based on the energy consumption optimization target models established in the above steps, the obtained environmental parameters are input into the n energy consumption optimization target models. The preset control parameter ranges are used as constraints, and heuristic algorithms are applied to solve each model to obtain the optimal solutions for the n sets of control parameters. The constraints are designed to meet human comfort requirements.

[0064] The heuristic algorithm includes, for example, at least one of the following: genetic algorithm, simulated annealing algorithm, particle swarm optimization algorithm, and ant colony optimization algorithm. By solving n optimization models (energy consumption optimization target models) to obtain n energy-saving optimization control directions, the energy-saving regulation control variable values ​​for all units in the region at the next time step are obtained, and units in the same group with similar conditions have the same regulation control variable values.

[0065] Because the Application Programming Interface (API) is used to input information and obtain return values,

[0066] The value returned by the API is encrypted, so the optimization model can be solved at one of the data providers (i.e., there is a corresponding server) (because the data provider has encryption tools that can decrypt the return value of the objective function).

[0067] This energy consumption optimization target model takes the outdoor temperature To and solar radiation rb as input for the next moment, and uses a heuristic algorithm to solve for the outlet water temperature Tout, compressor frequency fcom, and indoor temperature Ti of the air source heat pump unit. The output Tout and fcom are used as the control variables of the unit for the next moment (that is, by adjusting the outlet water temperature Tout and compressor frequency fcom in the next moment, the indoor temperature can be kept within the user's comfort range).

[0068] Step S130: The n sets of control parameters are sent to the n sets of air source heat pump units respectively to control the operation of the n sets of air source heat pump units respectively.

[0069] Based on the above step S120, n sets of control parameters are obtained. These n sets of control parameters are then used as the optimal energy-saving control direction and sent to the n air source heat pump units respectively to control their operation. The control parameters include: compressor frequency value, outlet water temperature, and / or indoor temperature. When this method is executed on the server, the server issues instructions, and the units operate according to these instructions.

[0070] To clearly illustrate the technical solution of the present invention, the execution flow of the control method for the air source heat pump unit provided by the present invention will be described below with reference to a specific embodiment.

[0071] Figure 4 This is a schematic diagram of a specific embodiment of the energy-saving control method for air source heat pump units provided by the present invention.

[0072] Step S1: Obtain the operating and installation data of all air source heat pump units in the target area;

[0073] Step S2: Divide all air source heat pump units in the target area into n groups of similar condition units.

[0074] All air source heat pump units are divided into multiple groups of similar condition devices, that is, into n groups of similar condition units. The data of each group of similar condition devices may be stored in k different servers, that is, the data of each group of similar condition devices (multiple devices) may have k data participants.

[0075] Step S3: Perform shared model training on units with similar conditions in each group to establish an energy consumption optimization target model for units with similar conditions in each group.

[0076] Train a shared model based on TEE for data of a single device with similar conditions (devices classified into a group with similar conditions), and use it as the target model for energy consumption optimization of that group of devices with similar conditions.

[0077] like Figure 3 As shown, the training process of the shared model for each group of devices with similar conditions includes:

[0078] S31, k data providers download and deploy encryption tools from the computing platform.

[0079] Due to privacy concerns regarding data (operational data and installation data), data from a single similar device exists on k servers belonging to k data participants and cannot be transferred between them.

[0080] S32, k data providers encrypt the data using encryption tools;

[0081] S33, the computing platform creates a trusted execution environment (TEE);

[0082] S34, k data providers send encrypted data to a trusted execution environment storage;

[0083] S35, The task initiator initiates the computation task;

[0084] S36, decrypt k data items and verify the authorization process;

[0085] S37. The computing platform trains the model based on the decrypted data to obtain the shared energy consumption model of the similar device, that is, the objective function of the optimization model (energy consumption optimization target model) corresponding to the similar device.

[0086] S38, transmits the encrypted calculation result;

[0087] S39, Destroy the trusted execution environment. (The destruction step is performed after the optimization model is solved; otherwise, the shared model cannot participate in the optimization model solution calculation.)

[0088] Step S4: Based on the energy consumption optimization target model obtained from the training, obtain the energy-saving optimization direction for each group of similar equipment.

[0089] The optimal measures for each group of similar condition devices can be obtained by training an optimization model for that group of similar condition devices.

[0090] Step S5: Issue energy-saving optimization control commands from each group of similar-condition units to the corresponding units.

[0091] Energy-saving optimization measures for each group of devices with similar conditions are centrally distributed to that group of devices. Specifically, for each group of devices with similar conditions in the target region, n shared models are trained, and n sets of optimized energy-saving measures are distributed.

[0092] Figure 5 An energy-saving optimization process for air-source heat pump units based on TEE-based shared machine learning according to the present invention is illustrated. For example... Figure 5 As shown:

[0093] 1) Selection of air source heat pump units under similar conditions;

[0094] The selection of air source heat pump units under similar conditions refers to units with similar performance and operating conditions. This is because two units operating under different conditions will have significantly different energy consumption characteristics. When constructing a migration network, operating data under identical conditions is more beneficial for training the federated model. Specific factors to consider when selecting units under similar conditions include: unit performance, climate zone, heat source installation location, and terminal equipment. The first factor is unit performance; units with different performance characteristics will have significantly different energy consumption models. The second factor is the climate zone in which the unit is located; units in extremely cold climate zones have higher energy consumption. The third factor is the heat source installation location; different locations, such as balconies, rooftops, and basements, are related to airflow. The fourth factor is the terminal equipment, such as underfloor heating, fan coil units, or radiators, which affects the unit's heating effect. Only when two units have the same performance, are located in the same climate zone, have comparable airflow at the heat source, and use the same terminal equipment can they be considered air source heat pump units under similar conditions.

[0095] 2) Feature selection;

[0096] Select features related to the energy consumption of air source heat pump units as model inputs, such as indoor and outdoor temperatures, compressor frequency, outlet water temperature, and solar radiation.

[0097] 3) Shared model training and optimization based on TEE;

[0098] Based on TEE (Transmission-Oriented Air Source Heat Pump) shared machine learning model training, air source heat pump units with similar conditions are collaboratively trained to form a shared model. This model is then used to predict the energy consumption of air source heat pump units with similar conditions.

[0099] 4) Obtain the objective function, build an energy-saving optimization model, solve it using a heuristic algorithm, and determine the direction of energy-saving optimization;

[0100] A complete energy-saving optimization model is built, and a heuristic algorithm is used to obtain the optimal solution, i.e., the energy-saving control variable. This guides the energy-saving direction of the unit. Specifically, the optimization model is as follows:

[0101]

[0102] Where E represents the shared energy consumption model, T i Indicates indoor temperature, T o Indicates outdoor temperature, f com T represents the compressor frequency. out Indicates the outlet water temperature, r b This represents the amount of solar radiation. Outdoor temperature To and solar radiation rb can be collected via web crawlers.

[0103] After constructing the objective function using a TEE-based shared model, the aforementioned inequalities regarding Tout, Ti, and fcom are used as constraints to obtain a constrained optimization model. These constraints are designed to meet human comfort requirements. Next, the optimization model is solved to obtain energy-saving control variable values ​​(e.g., the outlet water temperature and compressor frequency of the air source heat pump unit). Heuristic algorithms are used for solving this problem, such as genetic algorithms, ant colony optimization, particle swarm optimization, and simulated annealing. For example, genetic algorithms, due to their fast random search capabilities and potential parallelism, can effectively obtain solutions to the optimization model.

[0104] 5) Optimal instruction issuance;

[0105] Based on the aforementioned 4), the optimal energy-saving control direction is obtained, including the compressor frequency value and the outlet water temperature value. The server issues instructions, and the unit operates according to the instructions.

[0106] The present invention also provides a control device for an air source heat pump unit.

[0107] Figure 6 This is a structural block diagram of an embodiment of the control device for an air source heat pump unit provided by the present invention. Figure 6 As shown, the control device 100 of the air source heat pump unit includes: an acquisition unit 110, an output unit 120, and a distribution unit 130.

[0108] The acquisition unit 110 is used to acquire the environmental parameters of the target area at the next moment.

[0109] The environmental parameters may specifically include outdoor temperature and / or solar radiation. The outdoor temperature and / or solar radiation can be obtained using web crawling technology, for example, by crawling weather forecast websites to obtain the outdoor temperature and / or solar radiation for the next moment.

[0110] The output unit 120 is used to input the environmental parameters acquired by the acquisition unit into n pre-trained energy consumption optimization target models and output n sets of control parameters respectively.

[0111] The n sets of control parameters are used to control the n air source heat pump units at the next time step. Each of the n air source heat pump units includes two or more air source heat pump units with similar conditions. The n energy consumption optimization target models are trained on the n air source heat pump units respectively. That is, each group of air source heat pump units with similar conditions has a corresponding energy consumption optimization target model.

[0112] The control device 100 further includes a training unit (not shown), used to train the n energy consumption optimization target models in the following manner:

[0113] (1) Divide all air source heat pump units in the target area into similar conditions, so as to divide them into n groups of air source heat pump units with different similar conditions.

[0114] In one specific implementation, all air source heat pump units in the target area are divided into n groups of air source heat pump units with different similar conditions based on the unit's performance data, installation data, and local climate data.

[0115] The installation data may specifically include: heat source installation location and / or terminal type. For example, the similarity condition classification of air source heat pump units needs to consider factors such as: unit performance, climate zone, heat source installation location, and terminal. The first factor is unit performance; units with different performance characteristics have significant differences in energy consumption models. The second factor is the climate zone where the unit is located; units in extremely cold climate zones have higher energy consumption. The third factor is the heat source installation location; different locations, such as balconies, rooftops, and basements, are related to airflow. The fourth factor is the terminal, such as underfloor heating, fan coil units, or radiators, which affects the unit's heating effect. For example, two units can be considered similar air source heat pump units only if they have the same performance, are located in the same climate zone, have comparable airflow at the heat source, and have the same terminal. Each of the n groups of air source heat pump units has similar conditions. Each of the n groups of air source heat pump units includes two or more air source heat pump units with similar conditions.

[0116] (2) For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to the n groups of air source heat pump units.

[0117] Specifically, all air source heat pump units are divided into n groups of similar-condition devices (i.e., n groups of air source heat pump units with similar conditions). Data for each group of similar-condition devices (i.e., each group of air source heat pump units) may be stored on k different servers, meaning that data for each group of similar-condition devices (including two or more devices) may have k data participants. Specifically, the operational data for training the shared energy consumption model for each of the n groups of air source heat pump units is provided by k data providers. Each of the k data providers encrypts its provided operational data using a preset encryption tool and sends it to a trusted execution environment created by a preset computing platform for storage. The preset computing platform decrypts the encrypted operational data provided by the k data providers and performs shared model training to obtain the corresponding shared energy consumption model. Shared model training is then performed for each of the n groups of air source heat pump units, resulting in n shared energy consumption models corresponding to each of the n groups of air source heat pump units.

[0118] To address the issue of data silos in the operation of air source heat pumps from different manufacturers or even the same manufacturer in different regions, this invention employs a TEE-based shared model to obtain the objective function of the optimization model (energy consumption optimization target model). The optimization model is used for energy-saving optimization of computer groups (i.e., solving the corresponding optimization model).

[0119] For details on training the shared model, please refer to [link / reference]. Figure 3The training process of the shared model for each group of air source heat pump units with similar conditions is shown.

[0120] like Figure 3 As shown, the training process of the shared model for each group of devices with similar conditions includes:

[0121] S31, k data providers download and deploy encryption tools from the computing platform.

[0122] Due to privacy concerns regarding data (operational data and installation data), data from a single similar device exists on k servers belonging to k data participants and cannot be transferred between them.

[0123] S32, k data providers encrypt the data using encryption tools;

[0124] S33, the computing platform creates a trusted execution environment (TEE);

[0125] S34, k data providers send encrypted data to a trusted execution environment storage;

[0126] S35, The task initiator initiates the computation task;

[0127] S36, decrypt k data items and verify the authorization process;

[0128] S37. The computing platform trains the model based on the decrypted data to obtain the shared energy consumption model of the similar device, that is, the objective function of the optimization model corresponding to the similar device.

[0129] S38, transmits the encrypted calculation result;

[0130] S39, Destroy the trusted execution environment. (The destruction step is performed after the optimization model is solved; otherwise, the shared model cannot participate in the optimization model solution calculation.)

[0131] For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units. This model is used to predict the energy consumption of air source heat pump units with similar conditions. The training data for the shared energy consumption models selects features related to the energy consumption of the air source heat pump units as model inputs. The models are trained based on energy consumption, and the training data may include, for example, at least one of indoor temperature, outdoor temperature, compressor frequency, outlet water temperature, solar radiation, and power consumption.

[0132] Data providers encrypt and send data to a trusted computing platform. The platform decrypts the data and trains a model (which can be a relevant AI regression algorithm). Once certain conditions are met (reaching the required number of iterations or the error is less than a threshold), model training is complete, resulting in a shared energy consumption model for similar devices. Note that due to encryption, the trained model is not readily available. Data owners or others authorizing access to the shared model can input information and obtain return values ​​using an Application Programming Interface (API).

[0133] (3) The n shared energy consumption models are used as the objective functions of the energy consumption optimization target models of the n groups of air source heat pump units, and the energy consumption optimization target models corresponding to each group of air source heat pump units in the n groups of air source heat pump units are built respectively.

[0134] Due to encryption, the trained shared energy consumption model cannot be obtained. When the data owner or others authorize access to the shared energy consumption model, they can input information and obtain return values ​​using the Application Programming Interface (API). Let the encrypted shared energy consumption model be denoted as E. Since n shared energy consumption models (i.e., n objective functions) have been trained for each of the n groups of air-source heat pump units, n optimization models (i.e., n energy consumption optimization objective models) can be built accordingly. The complete energy consumption optimization objective models are shown below:

[0135]

[0136] Where E represents the shared energy consumption model, T i Indicates indoor temperature, T o Indicates outdoor temperature, f com T represents the compressor frequency. out Indicates the outlet water temperature, r b This represents the amount of solar radiation. The energy consumption optimization objective model limits the outlet water temperature T. out Compressor frequency f com Indoor temperature T i The control parameter range. Outdoor temperature To and solar radiation rb can be collected via web crawler. and The outlet water temperature T represents the water temperature T. out The maximum and minimum values; and These represent the compressor frequency f. com The maximum and minimum values, and These represent the indoor temperature T. i The maximum and minimum values.

[0137] Based on the aforementioned energy consumption optimization target model, the output unit 120 inputs the environmental parameters acquired by the acquisition unit 110 into the n energy consumption optimization target models, and uses the preset control parameter range as constraints to solve the n sets of optimal solutions using heuristic algorithms. The heuristic algorithms include, for example, at least one of genetic algorithms, simulated annealing algorithms, particle swarm optimization algorithms, and ant colony optimization algorithms. The constraints are used to meet human comfort requirements.

[0138] Solving n optimization models (energy consumption optimization target models) yields n energy-saving optimization control directions. This allows us to obtain the energy-saving regulation control variable values ​​for all units in the region at the next moment, and units in the same group with similar conditions have the same regulation control variable values.

[0139] Because the Application Programming Interface (API) is used to input information and obtain return values, and the values ​​returned by the API are encrypted, the optimization model can be solved at one of the data providers (i.e., there is a corresponding server) (because the data provider has encryption tools that can decrypt the return value of the objective function).

[0140] This energy consumption optimization target model takes the outdoor temperature To and solar radiation rb as input for the next moment, and uses a heuristic algorithm to solve for the outlet water temperature Tout, compressor frequency fcom, and indoor temperature Ti of the air source heat pump unit. The output Tout and fcom are used as the control variables of the unit for the next moment (that is, by adjusting the outlet water temperature Tout and compressor frequency fcom in the next moment, the indoor temperature can be kept within the user's comfort range).

[0141] The sending unit 130 is used to send the n sets of control parameters to the n sets of air source heat pump units respectively, so as to control the operation of the n sets of air source heat pump units respectively.

[0142] Based on the above, n sets of control parameters are obtained, and these n sets of control parameters are used as the optimal energy-saving control direction to be issued to the n air source heat pump units respectively, so as to control the operation of the n air source heat pump units respectively. The control parameters include: compressor frequency value, outlet water temperature, and / or indoor temperature. When the method is executed on the server, the server issues instructions, and the units operate according to the instructions.

[0143] The present invention also provides a storage medium corresponding to the control method of the air source heat pump unit, wherein a computer program is stored thereon, which, when executed by a processor, implements the steps of any of the aforementioned methods.

[0144] The present invention also provides a server corresponding to the control method of the air source heat pump unit, including a processor, a memory, and a computer program stored in the memory that can run on the processor, wherein the processor executes the program to implement the steps of any of the aforementioned methods.

[0145] The present invention also provides a server corresponding to the control device of the air source heat pump unit, including the control device of any of the aforementioned air source heat pump units.

[0146] Accordingly, the solution provided by the present invention, based on the technical solution of the present invention, obtains energy-saving optimization control parameters by using the same energy consumption optimization model for air source heat pump units with similar conditions. All units with similar conditions are trained together to form a model, which can effectively cope with small sample scenarios where operating data is limited in some regions or manufacturers.

[0147] According to the technical solution of the present invention, the shared machine learning construction method based on TEE obtains the objective function of the unit energy consumption optimization target model, effectively solving the problems of data silos and repeated model training. The entire calculation process is safe and reliable. By establishing a shared energy consumption model of air source heat pump units under the same conditions, the optimal control variables are solved using heuristic algorithms, which efficiently guides the energy-saving direction of the units, helps to improve user experience, save energy consumption, and at the same time reduces the computing pressure on the server and the problem of insufficient training data when modeling a single unit.

[0148] According to the technical solution of the present invention, it can replace a single optimization strategy for a single unit, use a shared machine learning modeling method based on TEE to solve the server operating pressure problem, and efficiently issue energy-saving optimization instructions to a large number of units.

[0149] The functions described herein can be implemented in hardware, software executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, the functions can be stored as one or more instructions or codes on or transmitted via a computer-readable medium. Other examples and embodiments are within the scope and spirit of this invention and the appended claims. For example, due to the nature of software, the functions described above can be implemented using software executed by a processor, hardware, firmware, hardwired, or any combination thereof. Furthermore, the functional units can be integrated into a single processing unit, or each unit can exist physically separately, or two or more units can be integrated into a single unit.

[0150] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0151] The units described as separate components may or may not be physically separate. Similarly, the components of the control device may or may not be physical units; they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0152] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0153] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A control method for an air source heat pump unit, characterized in that, include: Obtain the environmental parameters of the target area at the next moment; The acquired environmental parameters are input into n pre-trained energy consumption optimization target models, which output n sets of control parameters respectively. The n sets of control parameters are respectively sent to the n sets of air source heat pump units to control the operation of the n sets of air source heat pump units respectively; The n sets of control parameters are used to control the n sets of air source heat pump units at the next time step; each of the n sets of air source heat pump units has similar conditions; the n energy consumption optimization target models are trained for the n sets of air source heat pump units respectively. The n energy consumption optimization target models are trained in the following manner; The air source heat pump units in the target area are divided into n groups of air source heat pump units with different similar conditions, including: dividing the air source heat pump units in the target area into similar conditions based on the unit's performance data, installation data and local climate data; For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to the n groups of air source heat pump units. The n shared energy consumption models are used as the objective functions of the energy consumption optimization objective models of the n groups of air source heat pump units, and the energy consumption optimization objective models corresponding to each group of air source heat pump units in the n groups of air source heat pump units are built respectively. For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units, including: The operational data for training the shared energy consumption model for each of the n groups of air source heat pump units is provided by k data providers. The k data providers encrypt the provided running data for training the shared energy consumption model using a preset encryption tool and then send it to a trusted execution environment created by a preset computing platform for storage. The preset computing platform decrypts the encrypted running data provided by the k data providers for training the shared energy consumption model and then trains the shared model to obtain the corresponding shared energy consumption model.

2. The method according to claim 1, characterized in that, The acquired environmental parameters are input into n pre-trained energy consumption optimization target models, which output n sets of control parameters, including: After inputting the acquired environmental parameters into the n energy consumption optimization target models, the preset control parameter range is used as a constraint condition, and heuristic algorithms are used to solve the problem to obtain the optimal solution for the n sets of control parameters.

3. The method according to claim 1 or 2, characterized in that, The environmental parameters include: outdoor temperature and / or solar radiation. And / or, The control parameters include: compressor frequency, outlet water temperature, and / or indoor temperature.

4. A control device for an air source heat pump unit, characterized in that, include: The acquisition unit is used to acquire the environmental parameters of the target area at the next moment. The output unit is used to input the environmental parameters acquired by the acquisition unit into n pre-trained energy consumption optimization target models and output n sets of control parameters respectively. The sending unit is used to send the n sets of control parameters to the n sets of air source heat pump units respectively, so as to control the operation of the n sets of air source heat pump units respectively; The n sets of control parameters are used to control the n sets of air source heat pump units at the next time step; each of the n sets of air source heat pump units has similar conditions; the n energy consumption optimization target models are trained for the n sets of air source heat pump units respectively. It also includes: a training unit for training the n energy consumption optimization target models in the following manner: The air source heat pump units in the target area are divided into n groups of air source heat pump units with different similar conditions, including: dividing the air source heat pump units in the target area into similar conditions based on the unit's performance data, installation data and local climate data; For each of the n groups of air source heat pump units, a shared model is trained to obtain n shared energy consumption models corresponding to the n groups of air source heat pump units. The n shared energy consumption models are used as the objective functions of the energy consumption optimization objective models of the n groups of air source heat pump units, and the energy consumption optimization objective models corresponding to each group of air source heat pump units in the n groups of air source heat pump units are built respectively. The training unit performs shared model training for each of the n groups of air source heat pump units to obtain n shared energy consumption models corresponding to each of the n groups of air source heat pump units. This includes: the operating data for training the shared energy consumption models for each of the n groups of air source heat pump units is provided by k data providers; the k data providers encrypt the provided operating data for training the shared energy consumption models using a preset encryption tool and then send it to a trusted execution environment created by a preset computing platform for storage; the preset computing platform decrypts the encrypted operating data provided by the k data providers for training the shared energy consumption models and then performs shared model training to obtain the corresponding shared energy consumption model.

5. The apparatus according to claim 4, characterized in that, The output unit inputs the acquired environmental parameters into n pre-trained energy consumption optimization target models, and outputs n sets of control parameters, including: After inputting the acquired environmental parameters into the n energy consumption optimization target models, the preset control parameter range is used as a constraint condition, and heuristic algorithms are used to solve the problem to obtain the optimal solution for the n sets of control parameters.

6. A storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-3.

7. A server, characterized in that, It includes a processor, a memory, and a computer program stored in the memory that can run on the processor, wherein the processor executes the program to implement the steps of the method of any one of claims 1-3, and includes the control device of any one of claims 4-5.

Images