An air conditioning system

Abstract

This application relates to the field of home appliance control technology, and more particularly to an air conditioning system. The control method for the air conditioning system includes: acquiring thermal information of the building envelope, historical meteorological data, and actual outdoor temperature of the service area; wherein the historical meteorological data is meteorological data within a first preset time period; determining the hourly heat transfer offset and maximum heat load index of the building envelope of the service area for the next first preset time period based on the thermal information of the building envelope and the historical meteorological data; determining the real-time heat load index of the service area based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature; and determining the set temperature of the target terminal equipment as a first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range. This application can accurately determine the temperature setpoint, avoiding the risk of equipment overload and abnormal operating conditions of the terminal equipment.

Description

Technical Field

[0001] This application relates to the field of air conditioning technology, and more particularly to an air conditioning system. Background Technology

[0002] With the widespread adoption of central air conditioning systems in modern buildings, the demands of indoor occupants for thermal comfort in living and working environments continue to rise. Precisely matching indoor temperature to human needs has become one of the core objectives of HVAC operation and control.

[0003] In related technologies, a uniform temperature setpoint is typically used. However, this approach suffers from limitations because the thermal environment varies across different service areas of different terminal devices within a building, and even within the same service area at different times of the year. Consequently, the uniform temperature setpoint may fail to meet the cooling or heating needs of the service area. Alternatively, models based on historical operating energy consumption, indoor temperature and humidity, and load time-series data can be used to predict the temperature setpoint. However, this method relies on complete and continuous historical operating datasets, lacks scenario adaptability, and the predicted temperature setpoint is prone to exceeding the equipment's rated operating conditions and causing temperature control to exceed limits. This not only fails to accurately determine the temperature setpoint, affecting the stable operation of the air conditioning system, but also poses risks of equipment overload and abnormal operating conditions. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, or at least partially solve the above-mentioned technical problems, this application provides an air conditioning system.

[0005] This application provides an air conditioning system, including: an outdoor temperature sensor for collecting the actual outdoor temperature; multiple terminal devices, each terminal device having an indoor temperature sensor deployed thereon, the indoor temperature sensor being used to collect the actual indoor temperature of the service area where the terminal device is located; and a controller, the controller being communicatively connected to the outdoor temperature sensor, the indoor temperature sensor, and a target terminal device; the target terminal device being one of the multiple terminal devices.

[0006] The controller is configured to: acquire thermal information of the building envelope, historical meteorological data, and actual outdoor temperature of the service area; wherein the historical meteorological data is meteorological data within a first preset time period; based on the thermal information of the building envelope and the historical meteorological data, determine the hourly heat transfer offset and maximum heat load index of the building envelope of the service area for the next first preset time period, wherein the hourly heat transfer offset is used to indicate the inherent temperature rise characteristics of the service area indoors compared to outdoors without air conditioning intervention; based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature, determine the real-time heat load index of the service area, wherein the real-time heat load index is used to indicate the degree of heat intrusion into the service area at the current moment; and based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, determine the set temperature of the target terminal equipment as a first temperature value; wherein the first temperature value is used to instruct the target terminal equipment to perform temperature control based on the first temperature value.

[0007] This application also provides a control method for an air conditioning system, applicable to the air conditioning system in any of the above-mentioned technical solutions. The control method includes: acquiring thermal information of the building envelope, historical meteorological data, and actual outdoor temperature of the service area; wherein the historical meteorological data is meteorological data within a first preset time period; determining, based on the thermal information of the building envelope and the historical meteorological data, the hourly heat transfer offset and maximum heat load index of the building envelope of the service area for the next first preset time period, wherein the hourly heat transfer offset is used to indicate the inherent temperature rise characteristics of the service area indoors compared to the outdoors without air conditioning intervention; determining, based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature, the real-time heat load index of the service area, wherein the real-time heat load index is used to indicate the degree of heat intrusion into the service area at the current moment; and determining, based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, the set temperature of the target terminal device as a first temperature value; wherein the first temperature value is used to instruct the target terminal device to perform temperature control based on the first temperature value.

[0008] The air conditioning system provided in this application obtains hourly heat transfer offset and maximum heat load index through simulation, thereby determining the predicted heat load and maximum heat load for a first preset time period in the future. This facilitates the provision of a year-round benchmark for setting the temperature of terminal equipment, avoiding overload. Furthermore, since the hourly heat transfer offset is determined based on historical meteorological data and building envelope thermal information, the real-time heat load index calculated based on the hourly heat transfer offset can dynamically adapt to real-time meteorological and building envelope heat transfer conditions, making the determined real-time heat load more accurate and closely aligned with the actual operating state. Moreover, the first temperature setpoint of the terminal equipment is determined collaboratively based on the real-time heat load index, maximum heat load index, and set temperature allowable range, ensuring that the set temperature does not exceed the load limits of the equipment and building, and accurately determining the set temperature value corresponding to the terminal equipment, avoiding the problem of the set temperature exceeding the rated operating conditions of the target terminal equipment and temperature control exceeding limits. Attached Figure Description

[0009] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 The present application shows schematic diagrams of the air conditioning system in some embodiments; Figure 2 The diagram shows a structural block diagram of the controller in some embodiments of this application; Figure 3 The following diagram illustrates the air conditioning system deployment architecture in some embodiments of this application; Figure 4 The diagram shows a schematic representation of the controller in some embodiments of this application; Figure 5 A schematic diagram of all feature vectors in some embodiments of this application is shown; Figure 6 A progressive evolution diagram of temperature control by the controller in some embodiments of this application is shown; Figure 7 This illustration shows a staged evolution of the set temperature as the runtime changes in some embodiments of this application; Figure 8 A flowchart of a control method for an air conditioning system according to some embodiments of this application is shown; Figure 9A flowchart of another air conditioning system control method is shown in some embodiments of this application; Figure 10 A structural block diagram of the control device for an air conditioning system in some embodiments of this application is shown. Detailed Implementation

[0012] To make the objectives and implementation methods of this application clearer, the exemplary implementation methods of this application will be clearly and completely described below with reference to the accompanying drawings of the exemplary embodiments of this application. Obviously, the exemplary embodiments described are only some embodiments of this application, and not all embodiments.

[0013] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0014] The terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar or related objects or entities, and do not necessarily imply a specific order or sequence, unless otherwise specified. It should be understood that such terms are interchangeable where appropriate.

[0015] The terms “comprising” and “having”, and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device that includes a range of components is not necessarily limited to all of the components that are clearly listed, but may include other components that are not clearly listed or that are inherent to such product or device.

[0016] This application is applied to scenarios where temperature control is performed on different floors and in different areas of a public building's heating, ventilation, and air conditioning system.

[0017] In related technologies, a uniform temperature setpoint or a model-predicted temperature setpoint is typically used to determine the setpoint temperature for each terminal device within a building. Specifically, this uniform temperature setpoint method suffers from several drawbacks. Firstly, the uniform setpoint cannot meet the cooling or heating needs of the service area due to differences in orientation, floor level, functional zone, window-to-wall ratio, and building envelope parameters. Secondly, the uniform setpoint may fail to accurately determine the setpoint, as the temperature rise in a west-facing area due to afternoon sun exposure may be several times greater than in a north-facing area, even with the same setpoint, resulting in one side being overheated and the other undercooled. Thirdly, the model-predicted temperature setpoint method relies on complete and continuous historical operating datasets, which lacks adaptability to different scenarios. The predicted setpoint is prone to exceeding the rated operating conditions of the equipment and causing temperature control to exceed limits. This not only fails to accurately determine the setpoint and affects the stable operation of the air conditioning system but also poses risks of equipment overload and abnormal operating conditions.

[0018] To solve the above-mentioned technical problems, this application combines Figures 1 to 10 An air conditioning system and its control method are proposed.

[0019] Figure 1 The following are schematic diagrams of the air conditioning system in some embodiments of this application, such as... Figure 1 As shown, the air conditioning system 100 includes: an outdoor temperature sensor 110, an indoor temperature sensor 120, a target terminal device 130, and a controller 200. Optionally, the air conditioning system 100 may also include other terminal devices and other indoor temperature sensors.

[0020] Among them, the outdoor temperature sensor 110 is used to collect the actual outdoor temperature.

[0021] Indoor temperature sensor 120 is used to collect the actual indoor temperature of the service area where the terminal device is located. For example, indoor temperature sensor 120 is used to collect the actual indoor temperature of the service area where the target terminal device 130 is located.

[0022] The outdoor temperature sensor 110 and the indoor temperature sensor 120 collect data according to a control cycle. The control cycle can be set according to actual needs; for example, the control cycle can be 30 minutes.

[0023] In this embodiment, the controller 200 is communicatively connected to the outdoor temperature sensor 110, the indoor temperature sensor 120, and the target terminal device 130. The controller 200 can receive the actual outdoor temperature transmitted from the outdoor temperature sensor 110, the actual indoor temperature transmitted from the indoor temperature sensor 120, and can also send a set temperature to the target terminal device 130. Optionally, the controller 200 can also perform quality checks on the outdoor temperature sensor 110 and the indoor temperature sensor 120, including but not limited to: range detection, freeze detection, and temperature jump detection. When an anomaly is detected, the most recent valid value is used as the replacement value, and maintenance alarm information is generated.

[0024] Figure 2 The following are structural block diagrams of the controller in some embodiments of this application, such as... Figure 2 As shown, in Figure 2 In the controller 200, a passive thermal response feature library, an L1 physical protection layer, an L2 behavioral enhancement layer, a robustness assurance layer, and a time-series database are included. The passive thermal response feature library stores temperature simulation data. The L1 physical protection layer determines the physical protection temperature of the target end device; the L2 behavioral enhancement layer corrects the physical protection temperature; the robustness assurance layer ensures the robustness of the set temperature; and the time-series database stores sample feature vectors.

[0025] Figure 3 The following diagram illustrates the air conditioning system deployment architecture in some embodiments of this application, such as... Figure 3 As shown, in Figure 3 In this system, the field devices include an indoor temperature sensor 120, an outdoor temperature sensor 110, terminal devices 130, and a user control panel for user operation. The controller 200 utilizes an L1 physical protection layer and an L2 behavioral enhancement layer to determine the independent set temperature of each terminal device, and each terminal device performs temperature control based on its independent set temperature.

[0026] Figure 4 The following are schematic diagrams of the controller structure in some embodiments of this application, such as... Figure 4 As shown, the controller 200 includes a processor 201 and a memory 202 storing computer program instructions.

[0027] Specifically, the processor 201 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this disclosure.

[0028] Memory 202 may include a large-capacity storage device for information or instructions. For example, and not limitingly, memory 202 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 202 may include removable or non-removable (or fixed) media. Where appropriate, memory 202 may be internal or external to the integrated gateway device. In a particular embodiment, memory 202 is a non-volatile solid-state memory. In a particular embodiment, memory 202 includes read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (Electrically Programmable ROM, EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0029] The processor 201 reads and executes computer program instructions stored in the memory 202 to perform the steps of the control method of the air conditioning system 100 provided in this embodiment of the disclosure.

[0030] In one example, the controller 200 may also include a transceiver 203 and a bus 204. As shown in the figure, the processor 201, memory 202, and transceiver 203 are connected via the bus 204 and communicate with each other.

[0031] Bus 204 includes peripheral component interconnect (PCI) lines or extended industry standard architecture (EISA) buses. Bus 204 can be divided into address bus, data bus, control bus, etc.

[0032] The following describes the operation process of the air conditioning system 100 using the controller 200 executing the control method of the air conditioning system 100 as an example: When the air conditioning system is in the cold start period, the controller 200 is configured to: acquire thermal information of the building envelope, historical meteorological data and actual outdoor temperature of the service area; determine the hourly heat transfer offset and maximum heat load index of the building envelope of the service area for the first preset time period based on the thermal information of the building envelope and historical meteorological data; determine the real-time heat load index of the service area based on the actual outdoor temperature, hourly heat transfer offset and indoor reference temperature; and determine the set temperature of the target terminal device 130 as the first temperature value based on the real-time heat load index, maximum heat load index and set temperature allowable range value.

[0033] The cold start period is typically less than or equal to 7 days when the target terminal device is first started.

[0034] In this embodiment, the historical meteorological data refers to meteorological data within a first preset time period. The first preset time period can be set according to actual needs and is not limited. For example, the first preset time period can be one year. That is, the historical meteorological data is the hourly meteorological data of the service area throughout the year. For example, the historical meteorological data includes outdoor temperature, solar radiation, humidity, and wind speed.

[0035] In this embodiment, the first temperature value is used to instruct the target terminal device to perform temperature control based on the first temperature value. The first temperature value is also called the physical minimum temperature value.

[0036] In this embodiment, the building envelope thermal information includes building envelope thermal parameters, window parameters, and service area information for the target terminal equipment. The building envelope thermal parameters are various parameters of the building envelope components in terms of heat transfer, insulation, heat storage, airtightness, and shading. For example, building envelope thermal parameters include wall thermal conductivity, insulation layer thickness, airtightness rating, and shading coefficient. Window parameters include the window's heat transfer coefficient, shading coefficient, airtightness, and glass type. Service area information includes the area, volume, orientation, and floor height of the service area.

[0037] In this embodiment, the hourly heat transfer offset is used to indicate the inherent temperature rise characteristics of the service area indoors compared to the outdoors without air conditioning intervention.

[0038] In this embodiment, the maximum heat load index is the ultimate heat load index of the service area under the most severe operating conditions throughout the year. The maximum heat load index is used to indicate the upper limit of the heat transfer capacity of the building envelope under extreme operating conditions.

[0039] In this embodiment, the set temperature allowable range can be set according to actual needs and is not limited. For example, the set temperature operating range is 20°C to 28°C. The set temperature allowable range includes a maximum set temperature (e.g., a maximum set temperature of 28°C) and a minimum set temperature (e.g., a minimum set temperature of 20°C).

[0040] Understandably, by predicting the hourly heat transfer offset and the maximum heat load index, the predicted heat load and maximum heat load for the first preset time period can be determined, providing a year-round benchmark for setting the temperature of terminal equipment and avoiding overload. In addition, since the hourly heat transfer offset is determined based on historical meteorological data and building envelope thermal information, the real-time heat load index calculated based on the hourly heat transfer offset can dynamically adapt to real-time meteorological and building envelope heat transfer conditions, making the determined real-time heat load more accurate and closely match the actual operating conditions. Furthermore, the first temperature setpoint of the terminal equipment is determined collaboratively based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, ensuring that the set temperature does not exceed the load limits of the equipment and the building, and accurately determining the set temperature value corresponding to the terminal equipment, avoiding the problem of the set temperature exceeding the rated operating conditions of the equipment and temperature control exceeding the limits.

[0041] In some embodiments, when determining the hourly heat transfer offset and maximum heat load index of the building envelope of a service area for a future first preset duration based on the building envelope thermal information and historical meteorological data, the controller is configured to: simulate the building envelope thermal information and historical meteorological data to predict the hourly simulated indoor temperature and hourly simulated outdoor temperature of the service area within the future first preset duration; determine the maximum heat load index based on the hourly simulated indoor temperature and indoor reference temperature; and calculate the difference between the hourly simulated indoor temperature and the hourly simulated outdoor temperature to obtain the hourly heat transfer offset.

[0042] The hourly simulated indoor temperature indicates the indoor temperature of the service area for each hour within a first preset time period, as obtained from the simulation. The hourly simulated indoor temperature is also known as the hourly free-floating room temperature.

[0043] In this embodiment, the hourly simulated outdoor temperature is used to indicate the outdoor temperature for each hour within a future first preset time period obtained from the simulation.

[0044] In this embodiment, the indoor base temperature can be set according to the actual season without restriction. For example, the indoor base temperature can be set to 25°C in winter and 20°C in summer.

[0045] In one example, the controller 200 uses a building thermal simulation engine to build a digital twin model of the building and inputs the thermal information of the building envelope and historical meteorological data into the model. In the model, the air conditioning system is turned off, and hourly simulations of a first preset duration are performed. The hourly simulated indoor temperature and hourly simulated outdoor temperature of the service area are output. The maximum simulated indoor temperature is determined from multiple hourly simulated indoor temperatures, and the difference between the maximum simulated indoor temperature and the indoor reference temperature is calculated to obtain the maximum heat load index.

[0046] Understandably, by simulating the thermal information of the building envelope and historical meteorological data, the hourly simulated indoor and outdoor temperatures within a preset time period can be predicted. This allows for the determination of the maximum heat load index and the direct calculation of the hourly heat transfer offset based on the simulated indoor and outdoor temperature difference. This facilitates the subsequent calculation of the real-time heat load index based on the hourly heat transfer offset, providing a basis for accurately determining the set temperature. At the same time, the set temperature is constrained based on the maximum heat load index to avoid exceeding the target end equipment load.

[0047] In some embodiments, when determining the real-time heat load index of a service area based on the actual outdoor temperature, hourly heat transfer offset, and indoor reference temperature, the controller is configured to: calculate the sum between the actual outdoor temperature and the hourly heat transfer offset to obtain a first sum; calculate the difference between the first sum and the indoor reference temperature to obtain a first difference; and determine the maximum value between the preset value and the first difference as the real-time heat load index.

[0048] The default value is 0.

[0049] In this embodiment, the real-time heat load index is used to indicate the degree of heat intrusion into the service area at the current moment. The higher the real-time heat load index, the more severe the heat intrusion into the service area at the current moment, and the stronger the cooling compensation required.

[0050] Understandably, by using the actual outdoor temperature and the hourly heat transfer offset obtained in the simulation, the heat load of the service area can be accurately determined. This fully considers the temperature rise caused by the inherent heat transfer characteristics of the building envelope, and ensures that the index returns to zero under non-cooling conditions through preset numerical constraints. This effectively improves the accuracy of heat load judgment and the adaptability of operating conditions, and provides a reliable basis for the precise control of the cooling compensation intensity of the terminal equipment. In this way, the temperature can be accurately set to avoid over-cooling or under-cooling.

[0051] In some embodiments, when the set temperature of the target terminal device is determined to be a first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, the controller is configured to: calculate the ratio between the real-time heat load index and the maximum heat load index to obtain a first ratio; calculate the difference between the maximum set temperature and the minimum set temperature to obtain a second difference; calculate the product between the first ratio and the second difference to obtain a first regulating temperature; and calculate the difference between the maximum set temperature and the first regulating temperature to obtain the first temperature value.

[0052] The first regulated temperature satisfies:

[0053] in, This represents the first set temperature in the t-th hour; This represents the real-time heat load index for hour t. Indicates the maximum heat load index; This indicates the maximum set temperature. This indicates the minimum set temperature.

[0054] The first temperature value satisfies:

[0055] in, This represents the first temperature value in the t-th hour.

[0056] It should be noted that the stage of determining the first temperature value mentioned above is also called the first stage of temperature control (L1 stage).

[0057] Understandably, by linearly mapping the ratio of the real-time heat load index to the maximum heat load index to the set temperature allowable range, a physical minimum set temperature is obtained. This ensures that the set temperature is always within the set temperature allowable range and automatically adjusts the set temperature according to the real-time heat load differences in the service area. This provides a stable and reliable lower limit constraint for the subsequent intelligent optimization layer. At the same time, while ensuring that the indoor temperature is within a comfortable range, it fully releases the energy-saving potential based on heat load differences, improving the robustness, scenario adaptability, and overall operating energy efficiency of the air conditioning system.

[0058] In some embodiments, after determining the set temperature of the target terminal device as a first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range value, the controller is further configured to: acquire the first indoor actual temperature after a preset time period; detect whether the first indoor actual temperature is within a comfortable temperature range; if the first indoor actual temperature is not within a comfortable temperature range, determine the set temperature as a second temperature value based on the maximum or minimum set temperature value.

[0059] The second temperature value is used to instruct the target terminal device to perform temperature control based on the second temperature value.

[0060] In this embodiment, the preset time period can be set according to actual needs and is not limited. For example, the preset time period can be 15 minutes.

[0061] In this embodiment, the comfortable temperature range can be the same as or smaller than the set temperature range. For example, the comfortable temperature range is between 20°C and 25°C. The comfortable temperature range includes a maximum comfortable temperature and a minimum comfortable temperature.

[0062] In one example, if the actual temperature in the first room is less than the minimum comfortable temperature, then the second temperature value is determined to be the maximum set temperature; if the actual temperature in the first room is greater than the maximum comfortable temperature, then the second temperature value is determined to be the minimum set temperature.

[0063] It should be noted that the mechanism for determining the second temperature value is independent of the second stage (L2 stage) of temperature control. It has the highest priority and continues to be triggered until the actual indoor temperature returns to the comfortable temperature range.

[0064] Understandably, when the actual indoor temperature deviates from the comfortable temperature range, or becomes too low or too high, a second temperature value can be quickly determined directly according to the extreme value rule. This second temperature value determination mechanism has the highest priority, operates independently of the L2 stage of temperature control, and is not affected by the L2 stage temperature regulation strategy, thus avoiding regulation lag and oscillation problems caused by multi-stage control logic conflicts. Once triggered, it remains effective until the indoor temperature returns to the comfortable range, forcibly pulling the indoor temperature back to a reasonable and comfortable range, effectively avoiding discomfort caused by extreme deviations in room temperature, and improving environmental temperature control comfort.

[0065] Furthermore, when the target terminal device is in the rapid adaptation period, it enters the online ridge regression stage. Specifically, when the actual indoor temperature is within the comfortable temperature range, the controller is also configured to: normalize the real-time heat load index and the maximum heat load index to obtain the normalized heat load index; normalize the actual indoor temperature based on the hourly simulated indoor temperature to obtain the normalized indoor temperature; normalize the actual outdoor temperature based on the hourly simulated outdoor temperature to obtain the normalized outdoor temperature; generate a feature vector based on the preset bias term, the normalized heat load index, the normalized indoor temperature, the normalized outdoor temperature, and the time period; determine the predicted residual temperature value based on the feature vector and the preset weight matrix; calculate the sum of the residual temperature value and the first temperature value to obtain the third temperature value; and determine the set temperature as the third temperature value.

[0066] The rapid adaptation period is the time after the target terminal equipment has been started for a certain period of time. Typically, the rapid adaptation period is between 7 and 30 days after the air conditioning system is started.

[0067] In this embodiment, the preset bias term is a fixed value in the ridge regression model, and the preset bias term is 1.

[0068] In this embodiment, the time period is a One-Hot time period. The time period is one of multiple time periods throughout the day. These multiple time periods include a first time period, a second time period, a third time period, a fourth time period, and a fifth time period. Specifically, the first time period can be early morning; the second time period can be morning; the third time period can be noon; the fourth time period can be afternoon; and the fifth time period can be evening.

[0069] In this embodiment, the feature vector is:

[0070] in, Indicates the preset bias term; Indicates the normalized heat load index; This represents the normalized indoor temperature. This represents the normalized outdoor temperature. Indicates the first time period; Indicates the second time period; Indicates the third time period; Indicates the fourth time period; This indicates the fifth time period.

[0071] Understandably, the value of a given time period in the eigenvector differs from the values ​​of other time periods. For example, if the given time period is early morning, then... Set to 1; Set all to 0.

[0072] In this embodiment, the third temperature value is used to instruct the target terminal device to perform temperature control based on the third temperature value; the third temperature value is greater than or equal to the set minimum temperature value and less than or equal to the set maximum temperature value.

[0073] In this embodiment, the preset weight matrix is ​​the weight matrix corresponding to the target terminal device. For example, the weight matrix includes the weights corresponding to preset bias terms. Weights corresponding to the normalized heat load index Weights corresponding to normalized indoor temperature Weights corresponding to normalized outdoor temperatures The weight corresponding to the first time period The weight corresponding to the second time period The weight corresponding to the third time period The weight corresponding to the fourth time period The weight corresponding to the fifth time period .

[0074] Understandably, during routine automated operation, each end device maintains an independent weight memory matrix. After the physical layer's minimum temperature protection calculation is completed, the controller will retrieve the weight matrix specific to each end device.

[0075] In this embodiment, the predicted residual temperature value satisfies:

[0076] Understandably, normalizing key physical quantities such as heat load index and indoor / outdoor temperature avoids prediction bias caused by differences in data scale, thus improving the accuracy and reliability of the predicted residual temperature value. Incorporating the time period into the feature vector accurately captures users' temperature usage habits at different times, making the predicted third temperature value more consistent with users' dynamic behavior patterns. This avoids the shortcomings of a single static control strategy that cannot adapt to time-specific needs, improving the precision of temperature control. By summing the predicted residual temperature value with the first temperature value to obtain the third temperature value, and then determining it as the set temperature, it can effectively compensate for control biases caused by changes in environmental heat load and user behavior interference, making the set temperature closer to the user's actual comfort needs, reducing temperature fluctuations, and improving the stability of the indoor thermal environment and user comfort.

[0077] In some embodiments, when it is detected that the user actively performs temperature adjustment operation on the user control panel or sets control parameters through the platform, the preset weight matrix is ​​updated to obtain a new weight matrix.

[0078] Specifically, the user-set temperature corresponding to the temperature adjustment operation is obtained; the difference between the user-set temperature and the first temperature value is calculated to obtain the third difference; the difference between the third difference and the measured residual temperature value is calculated to obtain the prediction error; and the preset weight matrix is ​​updated using the stochastic gradient descent algorithm (SGD) to obtain the new weight matrix.

[0079] Wherein, any new weight in the new weight matrix satisfies:

[0080] in, Indicates the new weights; This represents the weights in the preset weight matrix; Indicates prediction error; The learning rate is usually expressed as a percentage of the total number of days. It is 0.1.

[0081] Optionally, This is related to the temperature direction regulation consistency ratio. For example, when the temperature direction regulation consistency ratio is greater than or equal to 0.8, The value is 0.1; when the temperature direction adjustment consistency ratio is less than 0.4, It is 0.1 The temperature direction adjustment consistency ratio can be determined based on the adjustment direction of the latest second adjustment temperature after a preset number of adjustments (e.g., 5 times).

[0082] Optionally, the controller 200 can also perform a decay calculation on the aforementioned weight matrix daily to obtain new weights. For example, the weight for the day is the product of yesterday's weight and the forgetting factor. The forgetting factor can be set according to the actual scenario, allowing the model to continuously track the latest user preferences and avoid old data dominating the learning results for a long time. For example, the forgetting factor can be 0.95 in service areas with high personnel flow; the forgetting factor can be 0.995 in service areas with fixed personnel.

[0083] Understandably, this ridge regression online learning method updates the weights and then recalculates the current new predicted residual temperature value with the new weights, thus obtaining a new third temperature value.

[0084] Furthermore, when the target terminal device is in the global optimization period, it enters the deep reinforcement learning stage. Specifically, after determining the set temperature as the third temperature value, the controller is also configured to: when the number of sample feature vectors is greater than a preset number, the running time of the air conditioning system is greater than a second preset time, and the variance of each feature component is less than a preset value within the third preset time, input the feature vector corresponding to the second preset time into the target policy network and output the second adjustment temperature; based on the second adjustment temperature and the first temperature value, determine the set temperature as the fourth temperature value.

[0085] The global optimization period is typically 30 days after the target terminal device is started.

[0086] In this embodiment, the sample feature vector includes a first type of sample feature vector and a second type of sample feature vector. The first type of sample feature vector indicates the user's acceptance of the current temperature control result. The second type of sample feature vector indicates the user's rejection of the current temperature control result.

[0087] In this embodiment, the preset quantity, the third preset duration, and the preset value can be set according to actual needs and are not limited. For example, the preset quantity is 500; the third preset duration is 7 days; and the preset value is 0.1.

[0088] In this embodiment, the second preset duration is the duration of the cumulative occupied period. The occupied period is the time during which the target terminal device is used by the user. The second preset duration can be set according to actual needs and is not limited. For example, the second preset duration is 30 days.

[0089] In this embodiment, the feature component is a term in the feature vector. The target policy network has the function of predicting the second regulation temperature.

[0090] In this embodiment, the fourth temperature value is used to instruct the target terminal device to perform temperature control based on the fourth temperature value.

[0091] Understandably, the second preset duration ensures that effective data accumulation and user habit learning are completed during the rapid adaptation phase before initiating the deep reinforcement learning phase. This avoids misjudgments and control fluctuations in the policy network due to insufficient data, guarantees the stability of indoor temperature during the transition phase, and prevents discomfort caused by sudden changes in control strategies. Furthermore, using the control results from the Ridge Regression phase as the initial state avoids the inefficiency of learning from scratch in reinforcement learning, achieving seamless integration and progressive optimization of the control strategy, and significantly shortening the convergence cycle of global optimization. The target policy network based on deep reinforcement learning can continuously learn complex factors such as long-term changes in user behavior, seasonal changes, and fluctuations in environmental thermal characteristics, dynamically adjusting the second regulating temperature. This allows the fourth temperature value to continuously adapt to various scenario changes during the global optimization period, exhibiting stronger environmental adaptability and robustness compared to fixed rules or linear models.

[0092] In some embodiments, before inputting the feature vector corresponding to the second preset duration into the target policy network and outputting the second temperature adjustment, the controller is further configured to: divide the multiple sample feature vectors into multiple feature vector groups; perform behavior cloning supervised training on the initial policy network based on multiple state vectors, and obtain an intermediate policy network after convergence; and perform supervised training on the intermediate policy network based on multiple state vectors, preset constraints, and preset reward functions to obtain the target policy network.

[0093] Each feature vector group corresponds to a state vector, and each feature vector group includes sample feature vectors from the past four consecutive periods.

[0094] Specifically, with the training objective of minimizing the imitation error, the initial policy network is trained under behavior cloning supervision based on multiple state vectors, and an intermediate policy network is obtained after convergence.

[0095] In this embodiment, minimizing the imitation error satisfies:

[0096] in, Represents the initial policy network; Represents the state vector; This indicates the third temperature value or the user-set temperature. This represents the minimization of the imitation error in the behavioral cloning loss.

[0097] Understandably, minimizing the behavior cloning loss can ensure the continuity of behavior after switching from the third temperature value to the fourth temperature value. In this way, the behavior cloning loss of the policy network in the sample feature vector gradually decreases and converges, resulting in the pre-trained intermediate policy network. The error between the output action of the intermediate policy network and the third temperature value or the user-set temperature in the sample feature vector is minimized, achieving behavior alignment with the original control strategy (ridge regression or user manual temperature adjustment).

[0098] In this embodiment, the preset reward function is a reward function composed of temperature comfort reward, energy saving reward, user coverage penalty, and safety boundary penalty.

[0099] The preset reward function satisfies:

[0100] in, This indicates the preset reward function; Indicates a reward for temperature comfort. This indicates the weight corresponding to the temperature comfort bonus; Indicates energy-saving rewards, This indicates the weight corresponding to the energy-saving reward; This indicates a user coverage penalty. This indicates the weight corresponding to the user coverage penalty; Indicates a penalty at the security boundary. This indicates the weight corresponding to the safety boundary penalty. The user coverage penalty is positively correlated with the frequency of user temperature adjustment operations within a single cycle; the safety boundary penalty is positively correlated with the distance between the fourth temperature value and the limit value of the set temperature range.

[0101] Specifically, satisfy:

[0102] in, The target temperature for the current time period (i.e., the midpoint of the user-set temperature or comfort temperature range).

[0103] Specifically, satisfy:

[0104] in, This indicates the second temperature setting. This indicates the first temperature value.

[0105] In this embodiment, the preset constraints may include the output range of the second regulated temperature output by the target policy network in each control cycle. Within this range, to prevent overly aggressive single-step decisions. Preset constraints may also include employing a proximal policy optimization (PPO) algorithm with additional divergence constraints to prevent excessively large update steps for the second regulating temperature.

[0106] It should be noted that before training the target policy network, sample feature vectors need to be obtained. Not all feature vectors are suitable as sample feature vectors. For example, feature vectors corresponding to temperature adjustments of less than 0.5°C or greater than 4°C, and feature vectors corresponding to temperature adjustments in opposite directions within 30 minutes, are considered invalid feature vectors. Figure 5 As shown, Figure 5 A schematic diagram of all feature vectors in some embodiments of this application is shown. Figure 5 In this model, all feature vectors include feature vectors of the first type of samples, feature vectors of the second type of samples, and invalid feature vectors. The process of determining sample feature vectors is described in detail below.

[0107] In some embodiments, during the temperature control process of the target terminal device based on a third temperature value, the controller is further configured to: detect whether temperature adjustment operation information is received; if no temperature adjustment operation information is received, and the time period in the feature vector is an occupied time period, then the feature vector is marked as a first-type sample feature vector. If temperature adjustment operation information is received, then the feature vector is marked as a second-type sample feature vector.

[0108] Among them, the temperature adjustment operation information is the operation information generated in response to the user's temperature adjustment operation. The first type of sample feature vector is collected only during the occupied period (such as during the day or when people are present) to avoid introducing invalid feature vectors during non-occupied periods.

[0109] Understandably, the absence of temperature adjustment by the user during the control cycle is considered a default acceptance of the current temperature control result, generating a first-type sample feature vector. The user's active temperature adjustment behavior is directly recorded as a second-type sample feature vector, which can accurately capture the user's clear temperature preference. The sample construction method combining the first-type and second-type sample feature vectors forms a complete user preference sample dataset, providing a sufficient and highly reliable training data source that meets the user's personalized comfort needs. At the same time, it reduces the interference of invalid samples on model training, improving learning convergence efficiency and control stability.

[0110] In some embodiments, during the process of temperature control by the target terminal device based on the fourth temperature value, the controller is further configured to: switch the set temperature from the fourth temperature value to the third temperature value if the user coverage frequency within the fourth preset time period is greater than a preset ratio, or if the fourth temperature value corresponding to each hour within the fourth preset time period does not belong to the comfortable temperature range, or if the first cumulative reward corresponding to the third preset time period is less than the second cumulative reward.

[0111] The user coverage frequency is used to indicate the frequency at which the user performs temperature adjustment operations.

[0112] In this embodiment, the first cumulative reward is a cumulative reward determined based on a preset reward function and a fourth temperature value. The second cumulative reward is a cumulative reward determined based on a preset reward function and a third temperature value.

[0113] In this embodiment, the fourth preset duration and preset ratio can be set according to actual needs and are not limited. For example, the fourth preset duration can be 24 hours, and the preset ratio can be 20%.

[0114] Understandably, by using the third temperature determined in the ridge regression stage as a benchmark and comprehensively considering three key indicators—user coverage frequency, action output variance, and cumulative reward—it can promptly and accurately identify performance degradation that occurs in the deep reinforcement learning stage, avoiding a decrease in control performance caused by policy model drift, overfitting, or environmental changes. When the deep reinforcement learning performance fails to meet the standards, it automatically switches back to the validated third temperature value determined in the ridge regression stage, avoiding problems such as large temperature fluctuations and frequent temperature adjustments by users caused by inaccurate output temperature settings in the deep learning stage, thus ensuring the basic comfort of the indoor thermal environment and the reliability of control.

[0115] Optionally, the air conditioning system can correspond to two independent model slots: one for cooling and one for heating, each containing its own weight matrix, policy network, and sample feature vector set. Seasonal switching is manually triggered by administrators or automatically executed on a preset date. During switching, the current season's model is frozen and the target season's model is loaded. If a historical model already exists for the target season, it continues learning from the model accumulated in the previous year, achieving model reuse.

[0116] Figure 6 This application illustrates a progressive evolution diagram of the controller performing temperature control in some embodiments, such as... Figure 6 As shown, in Figure 6In the process, the controller performs three types of quality checks on the sensor data: range, freeze, and jump, to ensure data validity. Then, the always-online L1 physical protection layer outputs the first temperature value as the basic control benchmark. After entering the L2 behavior enhancement layer, the system operates through a dual-path system of generational evolution and mutual backup. Online ridge regression outputs the third temperature value, and deep reinforcement learning outputs the fourth temperature value. The two use behavior cloning and degradation rollback mechanisms to achieve smooth transition and anomaly fallback. Finally, the set temperature is determined by combining the results of each level and sent to the target end device.

[0117] Figure 7 This paper illustrates a schematic diagram showing the stage evolution of the set temperature as a function of runtime in some embodiments of this application, such as... Figure 7 As shown, during the cold start period (0-7 days), only the physical baseline temperature is output, which is the first temperature value generated by the L1 physical baseline to ensure basic thermal comfort and operational safety. During the rapid adaptation period (7-30 days), the temperature value output by online ridge regression is superimposed on the physical baseline temperature to generate a set temperature (third temperature value) that is more in line with user habits. During the global optimization period (after 30 days): the strategy is switched to deep reinforcement learning, and the temperature value output by deep reinforcement learning is superimposed on the physical baseline temperature to obtain a fourth temperature value that takes into account comfort, energy consumption and user preferences, so as to achieve long-term global optimal control.

[0118] In some embodiments, Figure 8 A flowchart of a control method for an air conditioning system according to some embodiments of this application is shown, such as... Figure 8 As shown, the air conditioning system control method is applied to the air conditioning system in any of the above embodiments, and the air conditioning system control method includes: S801 obtains thermal information of building envelope, historical meteorological data and actual outdoor temperature of the service area.

[0119] Among them, historical meteorological data refers to meteorological data within the first preset time period.

[0120] S802, based on the thermal information of the building envelope and historical meteorological data, determines the hourly heat transfer offset and maximum heat load index of the building envelope in the service area for the first preset duration in the future.

[0121] S803 determines the real-time heat load index of the service area based on the actual outdoor temperature, hourly heat transfer offset, and indoor reference temperature.

[0122] S804 determines the set temperature of the target terminal equipment as the first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range. The first temperature value is used to instruct the target terminal device to perform temperature control based on the first temperature value.

[0123] In some embodiments, Figure 9 A flowchart of another control method for an air conditioning system according to some embodiments of this application is shown. Figure 9 In China, the control methods for air conditioning systems include: S901, calculate the ratio between the real-time heat load index and the maximum heat load index to obtain the first ratio.

[0124] S902, calculate the difference between the maximum set temperature and the minimum set temperature to obtain the second difference.

[0125] S903, calculate the product between the first ratio and the second difference to obtain the first adjustment temperature.

[0126] S904, calculate the difference between the maximum set temperature and the first adjusted temperature to obtain the first temperature value.

[0127] It should be noted that when the air conditioning system is controlled by the control device of the air conditioning system in the embodiments of this application, the specific implementation method is similar to the specific implementation method of the air conditioning system in any of the above embodiments of the present invention. Therefore, for a detailed exemplary description of the control process of the air conditioning system, please refer to the relevant description of the air conditioning system mentioned above. To reduce redundancy, it will not be repeated here.

[0128] In some embodiments, Figure 10 Structural block diagrams of the control device of an air conditioning system in some embodiments of this application are shown, such as... Figure 10 As shown, the control device for the air conditioning system is applied to the air conditioning system in any of the above embodiments. The control device 1000 for the air conditioning system includes: The acquisition module 1001 is used to acquire thermal information of the building envelope, historical meteorological data and the actual outdoor temperature of the service area; wherein the historical meteorological data is meteorological data within a first preset time period; The processing module 1002 is configured to: determine, based on the thermal information of the building envelope and the historical meteorological data, the hourly heat transfer offset and the maximum heat load index of the building envelope of the service area for the first preset time period in the future; the hourly heat transfer offset is used to indicate the inherent temperature rise characteristics of the service area indoors compared to outdoors without air conditioning intervention; determine the real-time heat load index of the service area based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature; the real-time heat load index is used to indicate the degree of heat intrusion into the service area at the current moment; and determine the set temperature of the target terminal device as a first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range value; wherein, the first temperature value is used to instruct the target terminal device to perform temperature control based on the first temperature value.

[0129] It should be noted that when the air conditioning system is controlled by the control device of the air conditioning system in the embodiments of this application, the specific implementation method is similar to the specific implementation method of the air conditioning system in any of the above embodiments of the present invention. Therefore, for a detailed exemplary description of the control process of the air conditioning system, please refer to the relevant description of the air conditioning system mentioned above. To reduce redundancy, it will not be repeated here.

[0130] This application discloses a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the methods described in the above embodiments.

[0131] This application discloses a computer program product, including a computer program, which, when executed by a processor, implements the methods described in the above embodiments.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0133] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the exemplary discussion above is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The embodiments were chosen and described to better explain the principles and practical applications, thereby enabling those skilled in the art to better utilize the embodiments and various different variations suitable for specific application considerations.

Claims

1. An air conditioning system, characterized in that, include: Outdoor temperature sensor, used to collect actual outdoor temperature; Multiple terminal devices, each of which is equipped with an indoor temperature sensor, which is used to collect the actual indoor temperature of the service area where the terminal device is located; A controller is communicatively connected to the outdoor temperature sensor, the indoor temperature sensor, and the target terminal device; the target terminal device is one of a plurality of terminal devices; the controller is configured to: The system acquires thermal information of the building envelope, historical meteorological data, and the actual outdoor temperature of the service area; wherein the historical meteorological data is meteorological data within a first preset time period. Based on the thermal information of the building envelope and the historical meteorological data, the hourly heat transfer offset and maximum heat load index of the building envelope of the service area for the first preset time period are determined. The hourly heat transfer offset is used to indicate the inherent temperature rise characteristics of the service area indoors compared to outdoors without air conditioning intervention. The real-time heat load index of the service area is determined based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature. The real-time heat load index is used to indicate the degree of heat intrusion into the service area at the current moment. Based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, the set temperature of the target terminal device is determined as a first temperature value; wherein, the first temperature value is used to instruct the target terminal device to perform temperature control based on the first temperature value; The step of determining the real-time heat load index of the service area based on the actual outdoor temperature, the hourly heat transfer offset, and the indoor reference temperature includes: Calculate the sum between the actual outdoor temperature and the hourly heat transfer offset to obtain a first sum. Calculate the difference between the first sum and the indoor reference temperature to obtain the first difference; The maximum value between the preset value and the first difference is determined as the real-time heat load index.

2. The air conditioning system according to claim 1, characterized in that, When determining the hourly heat transfer offset and maximum heat load index of the building envelope in the service area for the first preset duration based on the building envelope thermal information and the historical meteorological data, the controller is configured to: The thermal information of the building envelope and the historical meteorological data are simulated to predict the hourly simulated indoor temperature and hourly simulated outdoor temperature of the service area within the first preset time period in the future; wherein, the hourly simulated indoor temperature is used to indicate the simulated indoor temperature of the service area for each hour within the first preset time period in the future, and the hourly simulated outdoor temperature is used to indicate the simulated outdoor temperature for each hour within the first preset time period in the future. The maximum heat load index is determined based on the hourly simulated indoor temperature and the indoor reference temperature; The hourly simulated indoor temperature and the hourly simulated outdoor temperature are calculated to obtain the hourly heat transfer offset.

3. The air conditioning system according to claim 2, characterized in that, The set temperature allowable range includes the maximum set temperature value and the minimum set temperature value; When the set temperature of the target terminal device is determined to be a first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, the controller is configured to: Calculate the ratio between the real-time heat load index and the maximum heat load index to obtain a first ratio; Calculate the difference between the maximum set temperature and the minimum set temperature to obtain a second difference; The first regulating temperature is obtained by calculating the product of the first ratio and the second difference. The difference between the maximum set temperature and the first adjusted temperature is calculated to obtain the first temperature value.

4. The air conditioning system according to claim 3, characterized in that, After determining the set temperature of the target terminal device as the first temperature value based on the real-time heat load index, the maximum heat load index, and the set temperature allowable range, the controller is further configured to: The first actual indoor temperature is obtained after a preset time period; Detect whether the actual indoor temperature is within the comfortable temperature range; If the actual indoor temperature is not within the comfortable temperature range, the set temperature is determined as a second temperature value based on the set maximum or set minimum temperature value, wherein the second temperature value is used to instruct the target terminal device to perform temperature control based on the second temperature value.

5. The air conditioning system according to claim 4, characterized in that, When the first actual indoor temperature falls within the comfort temperature range, the controller is further configured to: The real-time heat load index and the maximum heat load index are normalized to obtain the normalized heat load index; Based on the hourly simulated indoor temperature, the actual indoor temperature is normalized to obtain the normalized indoor temperature. Based on the hourly simulated outdoor temperature, the actual outdoor temperature is normalized to obtain the normalized outdoor temperature. A feature vector is generated based on the preset bias term, the normalized heat load index, the normalized indoor temperature, the normalized outdoor temperature, and the time period. The predicted residual temperature value is determined based on the feature vector and the preset weight matrix; The sum of the residual temperature value and the first temperature value is calculated to obtain a third temperature value; wherein the third temperature value is greater than or equal to the set minimum temperature value and less than or equal to the set maximum temperature value; The set temperature is determined to be the third temperature value, wherein the third temperature value is used to instruct the target terminal device to perform temperature control based on the third temperature value.

6. The air conditioning system according to claim 5, characterized in that, After determining the set temperature as the third temperature value, the controller is further configured to: When the number of sample feature vectors is greater than a preset number, the running time of the air conditioning system is greater than a second preset time, and the variance of each feature component is less than a preset value within a third preset time, the feature vector corresponding to the second preset time is input into the target policy network, and the second temperature adjustment is output. The feature component is one of the feature vectors. Based on the second adjusted temperature and the first temperature value, the set temperature is determined to be a fourth temperature value; wherein, the fourth temperature value is used to instruct the target terminal device to perform temperature control based on the fourth temperature value.

7. The air conditioning system according to claim 6, characterized in that, Before inputting the feature vector corresponding to the second preset time period into the target policy network and outputting the second temperature adjustment, the controller is further configured as follows: The multiple sample feature vectors are divided into multiple feature vector groups, and each feature vector group corresponds to a state vector. Based on multiple state vectors, the initial policy network is trained under behavior cloning supervision, and an intermediate policy network is obtained after convergence. Based on multiple state vectors, preset constraints, and preset reward functions, the intermediate policy network is trained under supervision to obtain the target policy network. The preset reward function is a reward function composed of temperature comfort reward, energy saving reward, user coverage penalty, and safety boundary penalty.

8. The air conditioning system according to claim 7, characterized in that, The sample feature vector includes a first type of sample feature vector and a second type of sample feature vector; the first type of sample feature vector is used to indicate the user's acceptance of the current temperature control result; the second type of sample feature vector is used to indicate the user's non-acceptance of the current temperature control result. During the temperature control process performed by the target terminal device based on the third temperature value, the controller is further configured to: Detect whether temperature adjustment operation information has been received, wherein the temperature adjustment operation information is operation information generated in response to the user performing a temperature adjustment operation; If no temperature adjustment operation information is received, and the time period in the feature vector is an occupied time period, then the feature vector is marked as the first type of sample feature vector, where the occupied time period is the time period during which the target terminal device is used by the user; If the temperature adjustment operation information is received, the feature vector is marked as the second type of sample feature vector.

9. The air conditioning system according to claim 8, characterized in that, During the process of temperature control by the target terminal device based on the fourth temperature value, the controller is further configured to: If the user coverage frequency within the fourth preset duration is greater than a preset ratio, or if the fourth temperature value corresponding to each hour within the fourth preset duration does not belong to the comfortable temperature range, or if the first cumulative reward corresponding to the third preset duration is less than the second cumulative reward, the set temperature will be switched from the fourth temperature value to the third temperature value; wherein, the user coverage frequency is used to indicate the frequency at which the user performs the temperature adjustment operation; the first cumulative reward is a cumulative reward determined based on a preset reward function and the fourth temperature value, and the second cumulative reward is a cumulative reward determined based on the preset reward function and the third temperature value.

Images