A temperature control method for building space

Abstract

The present invention discloses a temperature control method for a building space, which belongs to the technical field of intelligent building temperature control, and comprises: S1, selecting a target temperature based on environmental status data, a building status range and a top-level target temperature strategy; S2, adjusting the building environmental temperature based on the target temperature and a bottom-level control and adjustment strategy to obtain real-time environmental temperature data; S3, judging whether the building environmental temperature adjustment is completed based on the real-time environmental temperature data and the target temperature; if the real-time environmental temperature data is equal to the target temperature, it is judged that the building environmental temperature adjustment is completed and the adjustment is ended; if the real-time environmental temperature data is not equal to the target temperature, it is judged that the building environmental temperature adjustment is not completed and S2 is executed; the present invention performs hierarchical control on the building environmental temperature through the top-level target temperature strategy and the bottom-level control and adjustment strategy, which can adapt to the complexity and variability of the building environment and significantly improve the temperature controllability of the building temperature control method.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent building temperature control, and in particular to a temperature control method for a building space. Background Art

[0002] With the development of artificial intelligence technology, many large buildings and shopping malls have realized the control of building space temperature based on artificial intelligence technology combined with sensors, chillers, air conditioners and other equipment. However, due to the limitations of sensor accuracy and the complexity and variability of the building environment, traditional building temperature control solutions that rely on precise sensor data and complex control prediction algorithms are difficult to take into account both energy saving and the comfort of people in the building space. Especially when the building environment is further complicated by the influence of weather, the temperature regulation of the corresponding building temperature control method is poor.

[0003] Chinese patent, publication number: CN118655792B, publication date: September 17, 2024, discloses a real-time monitoring system and intelligent adjustment method for building energy consumption based on the Internet of Things, including: a real-time monitoring module for identifying energy-consuming equipment in a building and collecting multiple parameter data inside the building in real time. The multiple parameter data inside the building include parameter signals of the internal environment of the building, relevant characteristic parameters of each energy-consuming equipment, and comfort data of users inside the building, and the collected data are classified and stored; an energy consumption analysis module for establishing an energy consumption analysis model and collecting real-time internal energy consumption data of the building. The relevant characteristic parameters of the equipment, various parameter signals of the internal environment of the building and the comfort data of the users in the building are input into the energy consumption analysis model to obtain the energy consumption analysis results; the energy consumption adjustment module is used to determine the operating power adjustment coefficient of the energy consumption equipment in the building according to the energy consumption analysis results of the energy consumption equipment in the building and the comfort data of the current users in the building, and judge whether the adjustment results of the energy consumption equipment in the building are optimal according to the comfort of the users in the building; however, this invention only considers energy saving, and does not consider the changes in building temperature relative to human comfort under complex environmental conditions, resulting in poor temperature controllability of this invention. Summary of the invention

[0004] The purpose of the present invention is to address the problem of poor temperature controllability of building temperature control methods caused by complex building space environments; a temperature control method for building space is proposed, which triggers the top-level target temperature strategy to select a target temperature when the building space temperature changes through environmental status data, and controls the building environment temperature based on the target temperature by running the bottom-level control adjustment strategy, and determines whether the adjustment process is terminated based on the real-time data of the ambient temperature and the target temperature. This method can adapt to the complexity and variability of the building environment and significantly improve the temperature controllability of the building temperature control method.

[0005] In a first aspect, a technical solution provided in an embodiment of the present invention is a temperature control method for a building space, comprising the following steps:

[0006] S1, selecting a target temperature based on environmental status data, building status range, and top floor target temperature strategy;

[0007] S2. Adjust the building ambient temperature based on the target temperature and the underlying control adjustment strategy to obtain real-time ambient temperature data;

[0008] S3. Based on the real-time ambient temperature data and the target temperature, determine whether the building ambient temperature adjustment is completed. If the real-time ambient temperature data is equal to the target temperature, determine that the building ambient temperature adjustment is completed and end the adjustment. If the real-time ambient temperature data is not equal to the target temperature, determine that the building ambient temperature adjustment is not completed and execute S2.

[0009] In this solution, whether the building ambient temperature needs to be adjusted is judged by the environmental status data and the building status range, so as to avoid the waste of building temperature regulation resources; when it is determined that the building ambient temperature needs to be adjusted, the target temperature is selected based on the environmental status data and the top-level target temperature strategy to obtain the target temperature suitable for the current building environment, and then the target temperature can be set as the policy target of the corresponding bottom-level control and regulation strategy to control the building ambient temperature, so that the building ambient temperature can be quickly increased or decreased, and can change with the changes in the complex environment in which the building is located, so as to achieve a stable relative relationship between the building ambient temperature and the complex environment in which the building is located, and significantly improve the temperature controllability of the building temperature control method; secondly, to ensure the effectiveness of building temperature regulation, whether the building ambient temperature regulation is completed is judged based on the real-time ambient temperature data and the target temperature in the building temperature regulation process. When it is not completed, that is, when the real-time ambient temperature data is not equal to the target temperature, the building ambient temperature can be continuously adjusted based on the target temperature and the bottom-level control and regulation strategy until the real-time ambient temperature data is equal to the target temperature. The regulation is terminated to avoid wasting temperature regulation resources after the building ambient temperature regulation is completed.

[0010] Preferably, before executing S1, the method further includes:

[0011] Obtain missing values ​​and outliers from the collected initial environmental status data based on statistical methods, and delete the outliers to obtain a missing data set;

[0012] Fill in the missing values ​​in the missing data set based on interpolation to obtain environmental status data;

[0013] The environmental status data at least includes environmental temperature data and environmental feedback data.

[0014] In this scheme, after collecting statistics of environmental status data based on statistical methods, we can find outliers and missing values ​​in the data that are different from the overall change trend. In order to facilitate unified processing, all outliers are deleted so that only missing values ​​exist in the environmental status data. Then, the interpolation method is used to fill in the corresponding data of the missing values ​​according to the change trend of the data to obtain complete and accurate environmental status data.

[0015] Preferably, in S1, the specific process of selecting the target temperature based on the environmental status data, the building status range and the top floor target temperature strategy is:

[0016] S11, extracting ambient temperature data and ambient feedback data from the ambient state data, and adjusting the ambient temperature data based on the ambient feedback data to obtain expected temperature data;

[0017] S12, judging whether to adjust the building ambient temperature based on the expected temperature data and the building status range, if the ambient temperature data is within the building status range, judging not to adjust the building ambient temperature and ending the adjustment, if the ambient temperature data is outside the building status range, judging to adjust the building ambient temperature and executing S13;

[0018] S13. Determine the target temperature by using the expected temperature data as input to the top-level target temperature strategy.

[0019] In this solution, the ambient temperature data and the ambient feedback data in the ambient status data are both data related to temperature. The difference is that the ambient temperature data is directly related to the temperature in the building environment, such as the overall average temperature of the area of ​​the building, and the ambient feedback data is directly related to the temperature change in the building environment, such as the temperature change caused by heat loss or heat gain caused by weather changes in the environment in which the building is located. The temperature change may be local or overall. In particular, when the temperature change occurs locally, the overall average temperature remains unchanged but the local temperature changes. By adjusting the ambient temperature data based on the ambient feedback data to obtain the expected temperature data, the refined control of the spatial temperature targets of each area inside the building can be achieved. Moreover, on the basis of the expected temperature data and combined with the preset building status range, accurate control of different areas of the building space under different weather conditions, different bottom temperature regulators, etc. can be achieved, thereby avoiding overall temperature regulation evolved due to local needs, and thus achieving effective energy saving.

[0020] Preferably, in S11, the specific process of adjusting the ambient temperature data based on the ambient feedback data to obtain the expected temperature data is:

[0021] The change difference is calculated based on the environmental feedback data and the environmental temperature data; when the change difference is equal to zero, the environmental temperature data is marked as the expected temperature data; when the change difference is greater than zero, the environmental feedback data is marked as the expected temperature data; when the change difference is less than zero, the environmental feedback data is marked as the expected temperature data.

[0022] In this solution, the difference in temperature reflecting the change in building environment temperature can be obtained by subtracting the ambient temperature data from the environmental feedback data, thereby accurately determining the changing trend of the building environment.

[0023] Preferably, in S13, the specific process of using the expected temperature data as the input of the top-level target temperature strategy to determine the target temperature is:

[0024] S131, using the expected temperature data as the adjustment target of the bottom temperature regulator corresponding to the top target temperature strategy to obtain the adjustment variable and the adjustment temperature, and calculating the adjustment reward according to the adjustment variable and the reward function;

[0025] S132, constructing a top-level tuple based on the adjusted temperature and the adjusted reward, establishing an experience pool based on the top-level tuple, and judging whether to update the expected temperature data according to the scale of the experience pool, when the scale of the experience pool is greater than the first scale threshold, judging to update the expected temperature data and executing S133, when the scale of the experience pool is less than or equal to the first scale threshold, judging not to update the expected temperature data and executing S131;

[0026] S133, randomly selecting a top-level tuple in the experience pool to obtain a random top-level tuple, and marking the adjusted temperature in the random top-level tuple as expected temperature data;

[0027] Simultaneously, the top-level loss difference is calculated based on the random top-level tuple and the loss function;

[0028] S134, adjusting the parameter combination of the top-layer target temperature strategy based on the top-layer loss difference, and judging whether to terminate the construction of the experience pool based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is less than or equal to the first iteration threshold, judging to continue to construct the experience pool and executing S131, if the number of adjustments of the parameter combination is greater than the first iteration threshold, judging to terminate the construction of the experience pool and executing S135;

[0029] S135. Sort the top-level loss values ​​to obtain the minimum top-level loss value, and mark the adjusted temperature in the top-level tuple corresponding to the minimum top-level loss value as the target temperature.

[0030] In this solution, if the building temperature is directly consistent with the ambient temperature, it is unreasonable from the perspective of energy saving or the comfort of people inside the building space. In order to make the target temperature of the building temperature adjustment process reasonable, the reward function and the adjustment variable are used to calculate the adjustment reward to quantify the effect of the temperature adjustment. The reward function also considers active reward variables, such as human comfort feedback reward, and passive reward variables, such as energy-saving effect reward, when calculating the reward function. The reward can be added in the update process of the adjusted temperature to make the adjusted temperature more and more reasonable, that is, a top-level tuple is constructed based on the adjusted temperature and the adjustment reward, and an experience pool is established based on the top-level tuple. The adjusted temperature is continuously updated through continuous cycles. New, but when the experience pool is no longer updated, the experience pool contains multiple top-level tuples, and each top-level tuple corresponds to an adjustment temperature. At this time, it is not certain which top-level tuple corresponds to the most reasonable target temperature. The loss function can be used to calculate the top-level loss value corresponding to the top-level tuple to judge the feasibility of the top-level tuple, and then find the most reasonable adjustment temperature that can be used as the target temperature. The parameter combination of the top-level target temperature strategy can also be adjusted based on the top-level loss value to achieve the optimization of the top-level target temperature strategy. On this basis, the expected temperature data is updated in combination with the randomly selected top-level tuple, which can avoid the high similarity of continuous expected temperature data during the experience pool update process and effectively avoid the target temperature selection from falling into the dilemma of local optimality.

[0031] Preferably, in S131, the reward function is specifically:

[0032] ;

[0033] In the formula, is the total reward value, is the active reward variable, is the passive reward variable.

[0034] Preferably, in S133, the loss function is specifically:

[0035] ;

[0036] In the formula, is the top layer loss difference, is the loss function, For the The input value corresponding to the top-level tuple randomly selected, is the input value corresponding to all top-level tuples as a whole, is the number of top-level tuples.

[0037] Preferably, after S1 is executed, environmental status data is acquired based on a preset real-time monitoring cycle and S1 is repeatedly executed.

[0038] Preferably, in S2, the specific process of adjusting the building ambient temperature based on the target temperature and the underlying control adjustment strategy to obtain the real-time data of the ambient temperature is:

[0039] S21, setting the target temperature as the policy target of the underlying control adjustment strategy, and selecting an adjustment action based on the policy target and the greedy algorithm;

[0040] S22, the corresponding bottom temperature regulator obtains the bottom temperature and bottom state in response to the adjustment action, constructs a bottom tuple based on the bottom temperature and the bottom state, and establishes a bottom experience pool based on the bottom tuple;

[0041] S23, judging whether to update the adjustment action based on the scale of the underlying experience pool and the second scale threshold, if the scale of the underlying experience pool is less than or equal to the second scale threshold, determining to update the adjustment action and executing S21, if the scale of the underlying experience pool is greater than the second scale threshold, determining not to update the adjustment action and executing S24;

[0042] S24, randomly selecting a bottom-level tuple in the bottom-level experience pool, setting the bottom-level state in the bottom-level tuple as the operating state of the bottom-level temperature regulator, and calculating the bottom-level loss difference according to the bottom-level tuple and the loss function;

[0043] S25, adjusting the parameter combination of the underlying control adjustment strategy based on the underlying loss difference, and judging whether to terminate the operation of the underlying control adjustment strategy based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is greater than the second iteration threshold, judging not to terminate the operation of the underlying control adjustment strategy and executing S21, if the number of adjustments of the parameter combination is less than or equal to the second iteration threshold, judging to terminate the operation of the underlying control adjustment strategy and executing S26;

[0044] S26. Sort the bottom layer loss differences to obtain the minimum bottom layer loss difference, and mark the bottom layer temperature in the bottom layer tuple corresponding to the minimum bottom layer loss value as the real-time data of the ambient temperature.

[0045] In this scheme, it can be found that the process of adjusting the building ambient temperature based on the target temperature and the bottom-level control and adjustment strategy to obtain real-time data of the ambient temperature is similar to the process of determining the target temperature by taking the expected temperature data as the input of the top-level target temperature strategy. The difference is that in the process of determining the target temperature by taking the expected temperature data as the input of the top-level target temperature strategy, the purpose is to obtain a reasonable target temperature, and the conditions that need to be considered are the energy-saving effect, the comfort of the human body in the building space, etc.; in the process of adjusting the building ambient temperature based on the target temperature and the bottom-level control and adjustment strategy to obtain real-time data of the ambient temperature, the purpose is to obtain reasonable actions of building temperature regulation, and the conditions that need to be considered are the bottom-level temperature regulators contained in the building area, the working status of the bottom-level temperature regulators, etc.

[0046] Preferably, the top-level target temperature strategy and the bottom-level control and regulation strategy are both set with priorities, and the priority of the top-level target temperature strategy is greater than the priority of the bottom-level control and regulation strategy.

[0047] In this solution, the priority of the top-level target temperature strategy is greater than the priority of the bottom-level control and adjustment strategy. The bottom-level control and adjustment strategy can be adjusted through the top-level target temperature strategy as soon as the target temperature changes due to changes in the building environment, thereby further improving the accuracy of building temperature control.

[0048] Beneficial effects of the present invention:

[0049] (1) The present application selects a target temperature based on environmental status data and a top-level target temperature strategy to obtain a target temperature suitable for the current building environment. The target temperature can then be set as the strategy target of the corresponding bottom-level control and adjustment strategy to control the building environment temperature, so that the building environment temperature can be quickly increased or decreased, and can change with changes in the complex environment in which the building is located, thereby achieving a stable relative relationship between the building environment temperature and the complex environment in which the building is located, and significantly improving the temperature controllability of the building temperature control method;

[0050] (2) When determining the target temperature of the building environment temperature, the present application selects the target temperature based on the environmental status data, the building status range and the top floor target temperature strategy. Through the corresponding reward function, the energy-saving effect and the impact of human comfort on the temperature are simultaneously considered, thereby improving the energy-saving effect in the temperature adjustment process and ensuring the temperature comfort for the human body. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0052] Figure 1 The figure is a flow chart of a temperature control method for a building space according to the present invention. DETAILED DESCRIPTION

[0053] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0054] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0055] Embodiment 1:

[0056] like Figure 1 As shown, this embodiment provides a temperature control method for a building space, comprising the following steps:

[0057] S0, that is, before executing S1, further includes:

[0058] Obtain missing values ​​and outliers from the collected initial environmental status data based on statistical methods, and delete the outliers to obtain a missing data set;

[0059] Fill in the missing values ​​in the missing data set based on interpolation to obtain environmental status data;

[0060] The environmental status data at least includes environmental temperature data and environmental feedback data.

[0061] In this embodiment, after collecting statistics of the environmental status data based on the statistical method, abnormal values ​​and missing values ​​that are different from the overall change trend can be found in the data. In order to facilitate unified processing, all abnormal values ​​are deleted so that only missing values ​​exist in the environmental status data. Then, the corresponding data of the missing values ​​are filled in according to the change trend of the data using the interpolation method to obtain complete and accurate environmental status data; the ambient temperature data is essentially the ambient temperature before the building space temperature changes, and the environmental feedback data is essentially the ambient temperature after the building space temperature changes. It should be noted that the change in the building space ambient temperature may be caused by adjustment, or it may be caused by factors such as weather changes.

[0062] S1. Select a target temperature based on the environmental status data, the building status range, and the top floor target temperature strategy.

[0063] S11, extracting ambient temperature data and ambient feedback data from the ambient state data, and adjusting the ambient temperature data based on the ambient feedback data to obtain expected temperature data;

[0064] The specific process of adjusting the ambient temperature data based on the ambient feedback data to obtain the expected temperature data is: calculating the change difference based on the ambient feedback data and the ambient temperature data; when the change difference is equal to zero, marking the ambient temperature data as the expected temperature data, when the change difference is greater than zero, marking the ambient feedback data as the expected temperature data, and when the change difference is less than zero, marking the ambient feedback data as the expected temperature data.

[0065] In this embodiment, the change difference can be obtained by subtracting the ambient temperature data from the environmental feedback data. When the environmental feedback data is equal to the ambient temperature data, the change difference is equal to zero, which proves that the corresponding building environment has not changed, and the ambient temperature data can be marked as expected temperature data. When the environmental feedback data is greater than the ambient temperature data, the change difference is greater than zero, which proves that the corresponding building environment temperature has risen, and the environmental feedback data can be marked as expected temperature data. When the environmental feedback data is less than the ambient temperature data, the change difference is less than zero, which proves that the corresponding building environment temperature has dropped, and the environmental feedback data can be marked as expected temperature data. At this point, real-time monitoring of the changing trend of the building environment can be achieved, and the expected temperature data can be updated as the building environment changes.

[0066] S12. Determine whether to adjust the building ambient temperature based on the expected temperature data and the building status range. If the ambient temperature data is within the building status range, determine not to adjust the building ambient temperature and end the adjustment. If the ambient temperature data is outside the building status range, determine to adjust the building ambient temperature and execute S13.

[0067] S13, using the expected temperature data as input to the top-level target temperature strategy to determine the target temperature;

[0068] S131, using the expected temperature data as the adjustment target of the bottom temperature regulator corresponding to the top target temperature strategy to obtain the adjustment variable and the adjustment temperature, and calculating the adjustment reward according to the adjustment variable and the reward function;

[0069] The reward function is specifically:

[0070] ;

[0071] In the formula, is the total reward value, is the active reward variable, is the passive reward variable;

[0072] S132, constructing a top-level tuple based on the adjusted temperature and the adjusted reward, establishing an experience pool based on the top-level tuple, and judging whether to update the expected temperature data according to the scale of the experience pool, when the scale of the experience pool is greater than the first scale threshold, judging to update the expected temperature data and executing S133, when the scale of the experience pool is less than or equal to the first scale threshold, judging not to update the expected temperature data and executing S131;

[0073] S133, randomly selecting a top-level tuple in the experience pool to obtain a random top-level tuple, and marking the adjusted temperature in the random top-level tuple as expected temperature data;

[0074] Simultaneously, the top-level loss difference is calculated based on the random top-level tuple and the loss function;

[0075] The loss function is specifically:

[0076] ;

[0077] In the formula, is the top layer loss difference, is the loss function, For the The input value corresponding to the top-level tuple randomly selected, is the input value corresponding to all top-level tuples as a whole, is the number of top-level tuples;

[0078] S134, adjusting the parameter combination of the top-layer target temperature strategy based on the top-layer loss difference, and judging whether to terminate the construction of the experience pool based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is less than or equal to the first iteration threshold, judging to continue to construct the experience pool and executing S131, if the number of adjustments of the parameter combination is greater than the first iteration threshold, judging to terminate the construction of the experience pool and executing S135;

[0079] S135. Sort the top-level loss values ​​to obtain the minimum top-level loss value, and mark the adjusted temperature in the top-level tuple corresponding to the minimum top-level loss value as the target temperature.

[0080] In this embodiment, before running the top-level target temperature strategy, it is necessary to initialize the top-level target temperature strategy, that is, initialize the parameter combination of the top-level target temperature strategy. The parameter combination of the top-level target temperature strategy can be adjusted in real time during the subsequent construction of the experience pool to adapt to the real-time changes in the building environment. At the same time, the process of determining the target temperature by the top-level target temperature strategy is a dynamic process of gradual optimization, which needs to cooperate with the bottom-level temperature regulator to calculate intermediate variables, that is, adjustment variables and adjustment temperature. In order to consider both the energy-saving effect and the temperature comfort of people in the building space during the temperature adjustment process, the adjustment reward calculated by the adjustment variable and the reward function is combined with the adjustment temperature to establish a top-level tuple. The top-level tuple can be expressed as , where is the building status corresponding to the current temperature, To select the target temperature, To adjust the rewards, In order to adjust the building state corresponding to the temperature, the top-level tuples are gathered together to form an experience pool. The size of the experience pool will increase with the increase in the number of top-level tuples. In order to make the final target temperature the best choice and to make the dynamic process of gradual optimization complete, top-level tuples will be continuously established until the size of the experience pool is greater than the first scale threshold. However, at this time, the top-level tuples in the experience pool are prone to fall into the dilemma of local optimality, resulting in the inability to obtain the most suitable target temperature for the current environment of the building. Therefore, the top-level tuples in the experience pool are randomly selected to obtain random top-level tuples, and the top-level loss difference is calculated based on the random top-level tuples and the loss function, so as to clarify the gap between the theoretical temperature corresponding to the top-level tuple in the experience pool and the expected temperature of the building environment through the top-level loss difference. This is because the top-level target temperature strategy The essence of the strategy is a neural network constructed based on a greedy strategy. The top-level target temperature strategy at least includes a Q network and a target network of deep reinforcement learning. The Q network corresponds to the theoretical temperature corresponding to the top-level tuple in the current experience pool, and the target network corresponds to the expected temperature of the building environment, that is, the subjective temperature desired to be obtained by adjusting the building temperature while considering both the energy-saving effect and the human comfort in the building space. The top-level loss difference can update the neural network parameters of the top-level target temperature strategy through back propagation, so that the theoretical temperature corresponding to the Q network and the expected temperature of the building environment corresponding to the target network gradually become consistent. In order to avoid overfitting or underfitting of the top-level target temperature strategy, the first iteration threshold is introduced to end the update iteration process of the top-level target temperature strategy at an appropriate time.

[0081] In one embodiment, after S1 is executed, environmental status data is acquired based on a preset real-time monitoring cycle and S1 is repeatedly executed.

[0082] S2. Adjust the building ambient temperature based on the target temperature and the underlying control adjustment strategy to obtain real-time ambient temperature data;

[0083] S21, setting the target temperature as the policy target of the underlying control adjustment strategy, and selecting an adjustment action based on the policy target and the greedy algorithm;

[0084] S22, the corresponding bottom temperature regulator obtains the bottom temperature and bottom state in response to the adjustment action, constructs a bottom tuple based on the bottom temperature and the bottom state, and establishes a bottom experience pool based on the bottom tuple;

[0085] S23, judging whether to update the adjustment action based on the scale of the underlying experience pool and the second scale threshold, if the scale of the underlying experience pool is less than or equal to the second scale threshold, determining to update the adjustment action and executing S21, if the scale of the underlying experience pool is greater than the second scale threshold, determining not to update the adjustment action and executing S24;

[0086] S24, randomly selecting a bottom-level tuple in the bottom-level experience pool, setting the bottom-level state in the bottom-level tuple as the operating state of the bottom-level temperature regulator, and calculating the bottom-level loss difference according to the bottom-level tuple and the loss function;

[0087] S25, adjusting the parameter combination of the underlying control adjustment strategy based on the underlying loss difference, and judging whether to terminate the operation of the underlying control adjustment strategy based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is greater than the second iteration threshold, judging not to terminate the operation of the underlying control adjustment strategy and executing S21, if the number of adjustments of the parameter combination is less than or equal to the second iteration threshold, judging to terminate the operation of the underlying control adjustment strategy and executing S26;

[0088] S26. Sort the bottom layer loss differences to obtain the minimum bottom layer loss difference, and mark the bottom layer temperature in the bottom layer tuple corresponding to the minimum bottom layer loss value as the real-time data of the ambient temperature.

[0089] In this embodiment, the bottom-level control and adjustment strategy is the same as the top-level target temperature strategy, both of which are neural networks constructed based on greedy strategies, and the basic principles are similar. The difference is that in the process of determining the target temperature by using the expected temperature data as the input of the top-level target temperature strategy, the purpose is to obtain a reasonable target temperature, and the conditions that need to be considered are energy-saving effects, human comfort in the building space, etc. In the process of adjusting the building ambient temperature based on the target temperature and the bottom-level control and adjustment strategy to obtain real-time ambient temperature data, the purpose is to obtain reasonable actions for building temperature regulation, and the conditions that need to be considered are the bottom-level temperature regulators contained in the building area, the working status of the bottom-level temperature regulators, etc.; the second scale threshold can be the same as the first scale threshold, the second iteration threshold can be the same as the first iteration threshold, and the bottom-level tuple can be approximately expressed as , where is the building status corresponding to the current temperature, The adjustment action of the bottom temperature regulator, To adjust the rewards, It is the corresponding building status after adjusting the temperature; in addition, when the bottom temperature regulator is an air conditioner, the working status of the bottom temperature regulator can be represented by data such as the number of normally working air conditioners, air conditioner cooling water temperature, air conditioner temperature, and air conditioner coolant flow.

[0090] S3. Based on the real-time ambient temperature data and the target temperature, determine whether the building ambient temperature adjustment is completed. If the real-time ambient temperature data is equal to the target temperature, determine that the building ambient temperature adjustment is completed and end the adjustment. If the real-time ambient temperature data is not equal to the target temperature, determine that the building ambient temperature adjustment is not completed and execute S2.

[0091] In one embodiment, the top-level target temperature strategy and the bottom-level control and regulation strategy are both set with priorities, and the priority of the top-level target temperature strategy is greater than the priority of the bottom-level control and regulation strategy.

[0092] In this embodiment, the priority of the top-level target temperature strategy is greater than the priority of the bottom-level control and adjustment strategy. When the building environment changes and causes the target temperature to change, the bottom-level control and adjustment strategy can be adjusted through the top-level target temperature strategy as soon as possible, thereby further improving the accuracy of building temperature control.

[0093] This embodiment has at least the following substantial effects:

[0094] (1) This embodiment selects a target temperature based on environmental status data and a top-level target temperature strategy to obtain a target temperature suitable for the current building environment. The target temperature can then be set as the policy target of the corresponding bottom-level control and adjustment strategy to control the building environment temperature, so that the building environment temperature can be quickly increased or decreased, and can change with the changes in the complex environment in which the building is located, thereby achieving a stable relative relationship between the building environment temperature and the complex environment in which the building is located, and significantly improving the temperature controllability of the building temperature control method;

[0095] (2) This embodiment selects the target temperature of the building environment temperature based on the environment status data, the building status range and the top floor target temperature strategy when determining the target temperature. The corresponding reward function simultaneously considers the energy-saving effect and the impact of human comfort on the temperature, thereby improving the energy-saving effect in the temperature adjustment process and ensuring the temperature comfort for the human body.

[0096] The above specific embodiments are preferred embodiments of the present invention, and are not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to the present specific embodiments. All equivalent changes made in accordance with the shape, structure, and method of the present invention are within the protection scope of the present invention.

Claims

1. A temperature control method for a building space, characterized in that: The method comprises the following steps: S1, selecting a target temperature based on environmental status data, building status range, and top floor target temperature strategy; S11, extracting ambient temperature data and ambient feedback data from the ambient state data, and adjusting the ambient temperature data based on the ambient feedback data to obtain expected temperature data; S12, judging whether to adjust the building ambient temperature based on the expected temperature data and the building status range, if the ambient temperature data is within the building status range, judging not to adjust the building ambient temperature and ending the adjustment, if the ambient temperature data is outside the building status range, judging to adjust the building ambient temperature and executing S13; S13, using the expected temperature data as input to the top-level target temperature strategy to determine the target temperature; S2. Adjust the building ambient temperature based on the target temperature and the underlying control adjustment strategy to obtain real-time ambient temperature data; S21, setting the target temperature as the policy target of the underlying control adjustment strategy, and selecting an adjustment action based on the policy target and the greedy algorithm; S22, the corresponding bottom temperature regulator obtains the bottom temperature and bottom state in response to the adjustment action, constructs a bottom tuple based on the bottom temperature and the bottom state, and establishes a bottom experience pool based on the bottom tuple; S23, judging whether to update the adjustment action based on the scale of the underlying experience pool and the second scale threshold, if the scale of the underlying experience pool is less than or equal to the second scale threshold, determining to update the adjustment action and executing S21, if the scale of the underlying experience pool is greater than the second scale threshold, determining not to update the adjustment action and executing S24; S24, randomly selecting a bottom-level tuple in the bottom-level experience pool, setting the bottom-level state in the bottom-level tuple as the operating state of the bottom-level temperature regulator, and calculating the bottom-level loss difference according to the bottom-level tuple and the loss function; S25, adjusting the parameter combination of the underlying control adjustment strategy based on the underlying loss difference, and judging whether to terminate the operation of the underlying control adjustment strategy based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is greater than the second iteration threshold, judging not to terminate the operation of the underlying control adjustment strategy and executing S21, if the number of adjustments of the parameter combination is less than or equal to the second iteration threshold, judging to terminate the operation of the underlying control adjustment strategy and executing S26; S26, sorting the bottom layer loss differences to obtain the minimum bottom layer loss difference, and marking the bottom layer temperature in the bottom layer tuple corresponding to the minimum bottom layer loss value as the real-time data of the ambient temperature; S3. Based on the real-time ambient temperature data and the target temperature, determine whether the building ambient temperature adjustment is completed. If the real-time ambient temperature data is equal to the target temperature, determine that the building ambient temperature adjustment is completed and end the adjustment. If the real-time ambient temperature data is not equal to the target temperature, determine that the building ambient temperature adjustment is not completed and execute S2.

2. A temperature control method for a building space according to claim 1, characterized in that: Before executing S1, the method further includes: Obtain missing values ​​and outliers from the collected initial environmental status data based on statistical methods, and delete the outliers to obtain a missing data set; Fill in the missing values ​​in the missing data set based on interpolation to obtain environmental status data; The environmental status data at least includes environmental temperature data and environmental feedback data.

3. A temperature control method for a building space according to claim 1, characterized in that: In S11, the specific process of adjusting the ambient temperature data based on the environmental feedback data to obtain the expected temperature data is: The change difference is calculated based on the environmental feedback data and the environmental temperature data; when the change difference is equal to zero, the environmental temperature data is marked as the expected temperature data; when the change difference is greater than zero, the environmental feedback data is marked as the expected temperature data; when the change difference is less than zero, the environmental feedback data is marked as the expected temperature data.

4. A building space temperature control method according to claim 1, characterized in that: In S13, the specific process of using the expected temperature data as the input of the top-level target temperature strategy to determine the target temperature is as follows: S131, using the expected temperature data as the adjustment target of the bottom temperature regulator corresponding to the top target temperature strategy to obtain the adjustment variable and the adjustment temperature, and calculating the adjustment reward according to the adjustment variable and the reward function; S132, constructing a top-level tuple based on the adjusted temperature and the adjusted reward, establishing an experience pool based on the top-level tuple, and judging whether to update the expected temperature data according to the scale of the experience pool, when the scale of the experience pool is greater than the first scale threshold, judging to update the expected temperature data and executing S133, when the scale of the experience pool is less than or equal to the first scale threshold, judging not to update the expected temperature data and executing S131; S133, randomly selecting a top-level tuple in the experience pool to obtain a random top-level tuple, and marking the adjusted temperature in the random top-level tuple as expected temperature data; Simultaneously, the top-level loss difference is calculated based on the random top-level tuple and the loss function; S134, adjusting the parameter combination of the top-layer target temperature strategy based on the top-layer loss difference, and judging whether to terminate the construction of the experience pool based on the number of adjustments of the parameter combination, if the number of adjustments of the parameter combination is less than or equal to the first iteration threshold, judging to continue to construct the experience pool and executing S131, if the number of adjustments of the parameter combination is greater than the first iteration threshold, judging to terminate the construction of the experience pool and executing S135; S135. Sort the top-level loss values ​​to obtain the minimum top-level loss value, and mark the adjusted temperature in the top-level tuple corresponding to the minimum top-level loss value as the target temperature.

5. A building space temperature control method according to claim 4, characterized in that: In S131, the reward function is specifically: R t =R c +R ec ; In the formula, R t is the total reward value, R c is the active reward variable, R ec is the passive reward variable.

6. A temperature control method for a building space according to claim 4, characterized in that: In S133, the loss function is specifically: In the formula, Loss is the top-level loss difference, MSE is the loss function, is the input value corresponding to the top-level tuple randomly selected for the i-th time, y is the input value corresponding to all top-level tuples as a whole, and n is the number of top-level tuples.

7. A building space temperature control method according to claim 1, characterized in that: After S1 is executed, environmental status data is acquired based on a preset real-time monitoring cycle and S1 is executed repeatedly.

8. A building space temperature control method according to claim 1, characterized in that: The top-level target temperature strategy and the bottom-level control and regulation strategy are both set with priorities, and the priority of the top-level target temperature strategy is greater than the priority of the bottom-level control and regulation strategy.

Images